‘v r. :rflwvwxcmzi '1": “7, 731;}, .5. 7.1 2... "‘J‘. 1' ’ I x 71 fl 12/1? 1773“,? wwi; "I '7‘ J 7" , '7 .‘ 4 33.797 7*‘14 ‘11:..7'7‘f7' .J' 7L -.7 .7124 . 73:34-73 283,2? 7%‘21‘ 1E . m- . ' u 27‘3"”; 3;; ragga,“ 73.111. ~ _. . 71' 34",! fin’y’vgfisw' i‘j’w‘ H ,..“3""£~J Hg. ,. 27,77. 4 I 13:. $171 dd 781 ‘ 1:? i :7 7 77734 '4 "7 I 43“, J J 7-71.14 1. 5,. l7 ‘1‘“ «7777737115375; 7 ”1 ‘1‘. ”£87“ I s!- , L:A~‘ 7 p_ 1 Q . 7777179651,»! .337"! ‘ “' " 7’. “(L-1%“?! -11§fi1%{7% 1.37.1457. ‘ A, fi'fi'fii’v 13173757”? gnu . w ,7 . . , ’I'R 9' #2“ e77. 557’“? ., T‘?’ 3%: , grv'pn *"7'7.,., 77': 777$ 15:77:: \7 7. kfhéT-Igiii‘ijfapr‘gkfigfiwfi '1‘?” .« "WW- M 7 5 77.. i I”*¢;?§§IT¢ ’7’}: ‘ ‘73.?! * 7 F‘ '7 fiffiififi“ ‘7 - 717:7?! 7", ’3' I <7 :77 " ”,1 .777 7777§%5;%$’*771‘LWI "'7 73$}. {£27197}! .‘L: 7.,7,.7.7,." +1: ' 7‘ U . ”"1." ‘ 7;“; j .77 777-77, 7.“: #7777771... .77 sfi7fifl7“:¢t&7’-Ilii‘ 77% . 1 77,7, 7,, . | , wzv.‘ . I‘m}- , III'JJIHI‘Ifig/g: 777‘ I"? a. 77.17.?77‘772377fi», 371:7? 7975:7777?!» 1:37;};1’flr‘4w‘7537 . 17'1‘3‘ 7’71 7‘47 ‘7' ' ' . ' 575,-! 'r‘I '\""7“‘ ‘7, "$747777 . 5%"7’1 'fiifiiwj‘fi- ‘ I ‘7 71?; IVJ‘VIU: 5.1!, fix" 775% we, .7. «:7: 77727171777: '72: ”mark? 27—: 7.7.4,, ’7wa 777:3): 7v~47'll‘ . ‘7 7,},‘77 7,,7 15‘ ' Rfigqhflf‘ ‘ J “7'73." ”rm"!!! 1:1, ’II"‘7“‘,?L' "‘" ,1“ _ 7777' 7:137“ 75“,” .ng‘ 411’ “7373!. '7 7 :flp‘fik" ""«"7"I*"!"""'” ‘7 L ' 71' fia’f'g‘h': “ ”7‘73!" 7 “-77.51.“ L‘L"‘7’( l.:17\:l 17:37.11}: 7’ gfg'fi’fiafl ligfiv, £371??? 7,7;.,'c,7~', 175,7; Z’Jfifigx‘wflfiia‘é; ,zgfi'ig,‘ If“. 7,. . #21,,“ 7. "f? ,1.“ ..,.,:g,1,,,,,',, ,5; ;,.ye7iflv,§,7y;.;,$,g;w ,7. #7,: m3 .‘ 77:77 7.;I.|7§77 7i .Jehsifilgffizut‘g‘m‘ltnui11'er ' "A ‘73? . I 22-33 7.37:; 7L1,.7‘I:,' .7“ 7.7 1,1,71‘770 7% Lfi-ggfit ;,;, 7 r; :7 7 .. I77. .,.,,7.,,,. 7w 7., ' ”41'7” 177%“ L “if” I 3‘3 ,,.,,, 7 ,_ 75:7,”! 77,77“. 7777; .- . 2?, 7+ "7‘77?!qu {7,}, t 1“.“ (r7, 7‘ 1,,an! $771,: 7;; 7 ex. .7773", 75 LI: 7; 1793:9777 .I 7""13‘ 15’" ‘ C 7‘7”” 7 7427’ “ 77174.!‘7713: 17:‘7’-‘i ‘71-}, ‘73:: 4 H 7."? ,' ..r. 151,775.37} 31.7w 7:71:57” ”247;“ 'I “’7‘..- “ 7 377‘ 3,,7‘.‘ 777M) 7 7:7,7' {'7}wa ” ‘7‘, 7,, ‘r v‘ -u—A '7‘: ‘ 7:77: ‘ 7, 71.717 ' ‘ .7, I » ‘ 77 . . 7‘ . '7’ ‘ .7,,7 1“,“ {‘7 7:623:77 ”15:” 75:7;y:?éi;l7{1‘777517x77»~"::’71'1r§,16‘i {777.7707 Lyrfldfé‘idr 3'21:-x V77, (.7, I? ,flfért, 7,7:‘J,I:7;{' 7 ‘- . 117‘. ‘17P —, w" ”77%;" (I; .61"; "“ ' “L“ 7‘7, 7;,7'7’1777 3'7“ ”(ikéu. 1L ,1: 7721!; Eflls'figgg ’ :7" .727 .7;ng flew: 7‘73. 74:35,;ch x 21. at 333527? ‘ ‘ " . 7.7‘ ,7, ' I" “7175771". $3??? "1'“? n :77 . ..,;,.7, 7.7 ' - ux€?.vi "‘ 1 i?” r“- Q 5.7.7733 V73} 7.77777 3, ,,,,..,,,J v. . 9 Ma 12131 ‘16:; .57., 7 r .1337: . . '“fi-‘v‘: .... 1" ... ' ' .,. .957. 7::7; . 7 . .. E _’.-.- .gggg 43,111, .ELIZV'Lr ‘ngg— _ ,_ , «Jami? . ,.. 1-: 317;, .3" mm: 31?? L "“amw'fiég‘gg ‘9 '. . . _ n 4‘ ’ ‘ M‘- :::7::. "mew :7; mm 17a .. I w 71v“! ,1: 233.7”: ' thaws \ o This is to certify that the thesis entitled ”"1 J .' "- .' ,. _ ‘1., -_-' , ‘ . . ‘ .gone--..-c au‘ murmur-21ml rauhms .ff—éhtjnc: the £193. 1:911. 3}“ Cattle in {ic‘ning presented bg '- , ‘ .\ ' __‘ ) '«~‘?‘C—~‘:' J. Lumen has been accepted towards fulfillment of the requirements for Thar). degree in “us. ', v; -; rig -'. D ' '—‘-'l on A-LL 1.. Major professor 0-169 "~—-- I 0‘ ‘ . l i. n /. \ \ '1 l \ I t l ‘ h ‘4 1 l .. I l I l I l l i I l. L\ t 1', ‘1 l / , l .‘ ., l i. l I ~. l 1 1- ‘ .- I \ — I O I I \ I / . \ ‘ '\ ‘ l h I ' ‘ I ,. . ‘Iv \ GENETIC AND ENVIRONMENTAL FACTORS AFFECTING THE RED DANISH CATTLE IN MICHIGAN BY Lester Joseph Cranek, Sr. A THESIS Submitted to the School of Graduate Studies of Michigan State College of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Dairy 1952 :‘ IHESIS ACKNOWLEDGMENTS The writer wishes to express his sincere appreciation to Dr. N. P. Ralston, Professor of Dairying, for his interest and con- structive suggestions throughout this study and for his aid in the critical reading and preparation of this manuscript. A large debt of gratitude is also owed Mr. G. E. Parsons, Assistant Professor of Dairy Extension, for his aid and many suggestions regarding the conduct of the survey this study entailed. Appreciation is also ex— pressed to Dr. Earl Weaver, Head, Dairy Department, Michigan State College, for the awards of the Michigan Artificial Breeders Association Fellowship and the Graduate Research Assistantship, and for the provision of the facilities required for the consummation of this study. - The author is indebted to Mr. A. C. Baltzer, Associate Professor of Dairy Extension, for his aid in placing at the disposal of the writer data invaluable to this study. Thanks are due Dr. J. A. Meiser, Jr., Assistant Professor of Dairying, for his assis— tance in the preparation of the photographic plates for the manu- script. Thanks are also extended to Dr. W. D. Baten, Professor of Mathematics, for his assistance with the statistical ‘analysis. 323774 iii The writer wishes to extend gratitude to his wife, whose never-ending encouragement has been invaluable. 7??? ' ? I‘d/"l ‘_-' l' ‘ :L’u' f H 2/” III 1‘. 5- ’ GENETIC AND ENVIRONMENTAL FACTORS AFFECTING THE RED DANISH CATTLE IN MICHIGAN By Lester Joseph Cranek, Sr. AN ABSTRACT Submitted to the School of Graduate Studies of Michigan State College of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Dairy Year 1952 Approved . I, ' f , _ A J' 'K' : 21 LESTER JOSEPH CRANEK, SR. ABSTRACT Analyses were made of all normal lactation records in the American Red Danish cattle population in Michigan prior to Aug- ust, 1951, to determine the effect on production characteristics of Red Danis sires when mated to cows of the various dairy breeds in a grading-up program. The progeny of forty-one herd sires were represented within the sire progeny groups there were avail- able 693 dams daughter comparisons. All lactation records of 2.70 days' duration or more were standardized to 305 day, 2X, mature equivalent basis by means of D.H.I.A. factors. A significant increase in milk and butterfat production of the Red Danish cross progeny over the foundation breeds was found. The mean milk and butterfat production for the foundation breeds and their graded-up progeny was 8,536 and 9,116 pounds of milk, and 354 and 372 pounds of butterfat, respectively. There was no evidence of heterosis among the various breeds and cross- bred animals. This increase was not attributed to an upward time trend in production. Estimates of repeatability of production rec— ords for the foundation breeds on an all herds and an intraherd basis were 0.54 and 0.36 milk production, and 0.66 and 0.37 for butterfat production, respectively. Those found for the Red Danish 2 LESTER JOSEPH CRANEK, SR. ABSTRACT progeny on an all herds and an intraherd basis were 0.52 and 0.26 for milk production, and 0.67 and 0.28 for butterfat production, respectively. The effect of mild inbreeding was analyzed by the intrasire regression of production on inbreeding. A nonsignificant decline of 23 pounds of milk and 0.3 pound of butterfat per l—percent in— crease in inbreeding was observed. The analysis showed that some sires withstand inbreeding, whereas others did not. The gets of inbred sires were no more uniform milk and butterfat production than those of noninbred sires. Two lethal defects, paralyzed hindquarters and mummifica- tion (a type of ankylosis), were found to be transmitted by the Red Danish sires. The observed frequency, of 2.2 and 1.4 per- cent for the paralyzed and mummified conditions, respectively, fitted the Mendelian inheritance ratio as simple recessive auto- somal characters. The frequency of the heterozygous carriers in the population for the paralyzed and mummified conditions ap- pears to be 25 and 11 percent, respectively. There appeared to be an increase of calf mortality and female sterility with each successive cross of Red Danish sires. LESTER JOSEPH CRANEK, SR. ABSTRACT The performance of herds sires as indicated by progeny and dam-daughter comparisons is presented for thirty-one sires having at least five dam-daughter pairs. Twenty-four sires in- creased production. When divided into sire family groups, it was found all families were alike in transmitting ability, as calculated by the ”Equal-Parent Index," which these data fitted. The influence of three nonhereditary factors, time trend and yearly environmental changes, month of calving, and length of calving interval on milk and butterfat production. Of the total observed variance for milk production, 2 and 3 percent was at- tributed to the month of calving and the length of calving interval, respectively, while 2.4, 3.2, and 6.7 percent of the total observed variance for butterfat production was attributed to the month of calving, length of calving interval, and year-to—year changes, respectively. TABLE OF CONTENTS GENERAL INTRODUCTION ..................... Brief History of the American Red Danish Cattle Development of the Red Danish Milkrace in Denmark .............................. Importation to the United States .............. Entrance into Michigan .................... Development of the Red Danish Cattle in Michigan .............................. Origin of the American Red Danish Cattle Association ............................ GENERAL OBJECTIVES ....................... GENERAL INVESTIGATIONAL PROCEDURE ......... Source of Data ............................ Placing of Data on IBM Cards ................. Punching of Data from Worksheets ............ Punching of Inbreeding Coefficients ........... Standardization of Records for Age and Length of Records ....................... 20 20 24 28 Calculation of Average Lifetime Production . Calculation of Total Sums of Squares ..... Calculation of the Between Sums of Squares Calculation of Regression of Production on Inbreeding ........................ Calculations of Sire Evaluation and Estimates of Heritability ..................... Methods Used in Deriving Coefficients of Inbreeding .......................... Females ......................... AVERAGE WLK AND BUTTERFAT PRODUCTION, REPEATABILITY, AND HERITABILITY OF COWS RESULTING FROM THREE SUCCESSIVE CROSSES OF PUREBRED RED DANISH SIRES .......... Review of Lite rature .................. Grading—Up Effect ....................... Introduction ..................... ..... ..... ..... 00000 Repeatability of Milk, Butterfat, and Butterfat ' Percentage ....................... Introduction ..................... Page 32 35 38 40 44 47 47 54 59 59 59 59 61 69 73 73 Calculation of Average Lifetime Production ...... Calculation of Total Sums of Squares ..... Calculation of the Between Sums of Squares Calculation of Regression of Production on Inbreeding ........................ Calculations of Sire Evaluation and Estimates of Heritability ..................... Methods Used in Deriving Coefficients of Inbreeding ............................... Females ......................... AVERAGE MLK AND BUTTERFAT PRODUCTION, REPEATABILITY, AND HERITABILITY OF COWS RESULTING FROM THREE SUCCESSIVE CROSSES OF PUREBRED RED DANISH SIRES .......... Review of Lite rature .................. Grading—Up Effect ....................... Introduction ..................... 00000 Repeatability of Milk, Butterfat, and Butterfat ' Percentage ....................... Introduction ..................... Page 32 35 38 40 44 47 47 54 59 59 59 59 61 69 73 73 vi Page Methods for Estimating Repeatability ........ 74 Partial correlation ................... 74 Analysis of variance .................. 75 Intraclass correlation ................. 76 Estimates of Repeatability ................ 76 Heritability of Production of Milk, Butterfat, and Butterfat Percentage ................... 77 Introduction .......................... 77 Methods for Estimating Heritability ......... 86 Isogenic line method .................. 87 Regression of offspring on mid—parent ..... 88 Correlation of parent and offspring ....... 89 Intrasire dam—daughter regression or correlation ........................ 90 Half-sib correlation .................. 91 Full—sib correlation .................. 91 Estimates of Heritability ................. 92 Investigational Procedure .................. '. . 1oo Grading—Up Effect ....................... 100 Repeatability Estimates .................... 101 Heritability E stimate s ..................... 10 5 vii Page Investigational Results ...................... 106 Grading—Up Effect ....................... 106 Repeatability Estimates .................... 129 Heritability Estimates ..................... 136 Discussion and Conclusion .................... 139 Grading—Up Effect ....................... 139 Repeatability Estimates .................... 142 Heritability Estimates ..................... 149 Summary ................................ 152 EVALUATION OF THE PRODUCTION- TRANSMITTING ABILITY OF THE PUREBRED RED DANISH SIRES USED IN MICHIGAN ........... 156 Review of Literature ....................... 156 Introduction ............................ 1 5 6 Methods for Evaluating A Sire's Transmitting Ability ............................... 157 Minimum Number of Progeny for Sire Evaluation ............................. 175 Daughters ........................... 175 Sons ............................... 180 Investigational Procedure .................... 181 Investigational Re sults ...................... 1 81 Investigational Results ...................... Grading—Up Effect ....................... Repeatability Estimates .................... Heritability Estimates ..................... Discussion and Conclusion .................... Grading-Up Effect ....................... Repeatability Estimates .................... Heritability Estimates ..................... Summary ................................ EVALUATION OF THE PRODUCTION- TRANSMITTING ABILITY OF THE PUREBRED RED DANISH SIRES USED IN MICHIGAN ........... Review of Literature ....................... Introduction ............................ Methods for Evaluating A Sire's Transmitting Ability ............................... Minimum Number of Progeny for Sire Evaluation ............................. Inve stigational P roc e du re .................... Inve stigational Re 3 ults ...................... vii Page 106 106 129 136 139 139 142 149 152 156 156 156 157 175 175 180 181 181 l8- viii Page Discussion and Conclusions ................... 186 Summary ................................ 190 THE EFFECTS OF MILD INBREEDING ON MILK AND BUTTERFAT PRODUCTION AND THE OCCURRENCE OF DEFECTS ................ 192 Review of Literature ....................... 192 Introduction ............................ 192 Inbreeding of Laboratory and Farm Animals ..... 193 Inbreeding of Dairy Cattle .................. 195 Production ........................... 195 Birth Weight ......................... 199 Size and Weight at Two Years and Over ...... 202 Type ............................... 204 Fertility ............................ 206 Mortality of Calves .................... 207 Occurrence of Lethal Defects ............. 208 Investigational Procedure .................... 209 Investigational Results ...................... 212 The Effects of Mild Inbreeding on Milk and Butterfat Production ................... 212 Lethal Defects .......................... 220 Other Effects ......................... Calf Mortality ...................... Sterility .......................... Discus sion and Conclusion .................. Effects of Mild Inbreeding on Milk and Butterfat Production ................ Lethal Defects ........................ Other Effects ......................... Summary .............................. THE INFLUENCE OF THREE NONHEREDITARY FACTORS: TIME TREND AND YEARLY ENVIRONMENTAL CHANGES, MONTH OF CALVING, AND CALVING INTERVAL ON MILK AND BUTTERFAT PRODUCTION ............... Review of Literature ..................... Introduction .......................... Year—to—Year Variation and Time Trend ...... Influence of Month of Calving on Milk and Butterfat Production and Butterfat Per- centage ............................. Effect of Season of Calving ............. Effect of Low Temperature ............. Effect of High Temperature ............. Page 230 230 232 232 232 235 238 238 240 240 240 241 -242 243 248 250 Page Influence of Calving Interval on Milk and Butterfat Production ...................... 255 Investigational Procedure .................... 259 Investigational Results ...................... 262 Time Trend and Yearly Environmental Changes . . . 262 Effect of the Month of Calving on Milk and Butterfat Production ...................... 270 Effect of Calving Interval on Milk and Butterfat Production ...................... 279 Discussion and Conclusions ................... 286 Time Trend and Yearly Environmental Changes . . . 286 Effect of Month of Calving on Milk and Butterfat Production ...................... 287 Effect of Length of Calving Interval on Milk and Butterfat Production ............... 292 Summary ................................ 293 LITERATURE CITED ......................... 295 APPENDIX ................................ 317 GENERAL INTRODUCTION An intense search for superior animals by the United States ‘ Department of Agriculture in the 1930's led to the importation of two bulls and twenty cows of the Red Danish Milkrace breed from Denmark. In order to more thoroughly evaluate the genetic com- position of the animals of this importation, it became necessary to test a high percentage of the ‘males. Cooperating dairymen in Michigan were enlisted to prove these bulls under the supervision of Mr. A. C. Baltzcr of Michigan State College. The characteristics of these Red Dane progeny were such that they satisfied an apparent desire of these cooperating dairy— men to the extent that on January 16, 1948, the American Red Danish Cattle Association was formed. An open herdbook was es—- tablished, and third- and fourth-generation females and males, respectively, would become eligible for registration, providing the foundation cows; and all succeeding females, were production tested. Anyone breeding Red Danish cattle for the purpose of final regis- tration of the third and succeeding generations must comply with the birth—reporting and production—testing regulations. GENERAL INTRODUCTION An intense search for superior animals by the United States - Department of Agriculture in the 1930's led to the importation of two bulls and twenty cows of the Red Danish Milkrace breed from Denmark. In order to more thoroughly evaluate the genetic com- position of the animals of this importation, it became necessary to test a high percentage of the males. Cooperating dairymen in Michigan were enlisted to prove these bulls under the supervision of Mr. A. C. Baltzer of Michigan State College. The characteristics of these Red Dane progeny were such that they satisfied an apparent desire of these cooperating dairy— men to the extent that on January 16, 1948, the American Red Danish Cattle Association was formed. An open herdbook was es— tablished, and third- and fourth-generation females and males, rESpectively, would become eligible for registration, providing the foundation cows; and all succeeding females, were production tested. Anyone breeding Red Danish cattle for the purpose of final regis- tranion of the third and succeeding generations must comply with the birth-reporting and production-testing regulations. 2 The establishment of a new breed at this stage in our knowl- edge of animal breeding offers an entirely new approach to the rapid development of a breed that will possess significant economic char- acteristics. Such an assertion seems warranted in view of the vast accumulation of breeding knowledge. Little time has elapsed to allow unsound fads and fancies and unintelligent promotional activities and the development of policies to restrict the use of fundamental objectives outlined in their program and keep abreast of current breeding information, they can avoid many of the pit- falls which have delayed progress in some of our other breeds. This investigation was initiated to establish the trend of the essential characteristics in the development of the American Red Danish cattle up to the present. The data presented in this study should materially aid in breed progress, as it was designed to de— termine the repeatability and heritability of production characters, to find the superior breeding animals, to observe the frequency of undesirable defects, and to emphasize some of the major environ- mental factors responsible for phenotypic variation. fl“— Brief History of the American Red Danish Cattle Development of the Red Danish Milkrace in Denmark The early history of the Red Danish Milkrace is veiled in obscurity, as is equally true with the history of the other dairy breeds of the world. I The cattle found in Denmark at the end of the period 1778 were of no distinct color, except those in the northern part of Jut- land, which were mostly black, or white and black. In the southern and eastern part of Denmark, in which regions the Red Danish dairy breed was to later develop, the cattle consisted of an admix— ture of color and type. These cattle were of small to average size, were horned and generally spotted in color, red and white, yellow, or brindle, and a little tendency to showing black. It is certain that the native cattle of the western part of Jutland were of a more dual-purpose type and presented a more fixed type at the beginning of the nineteenth century than did the cattle of any other part of Denmark (Anthony, 1951). From 1800 to 1840, several types of cattle were brought to the islands of Funen, Sealand, and Lolland-Falster. This infusion 01' cattle consisted of Sleswig and Holstein cattle, some Jutland cattle, traces of Ayrshire, Swiss and Tyrol blendings, and some Angler cattle. The Sleswig cattle (Ballum and north Sleswig) pre- dominated, and the red color was the most predominant. This movement occurred because of large quantities of surplus feeds from the distilleries and breweries on the islands of Funen and Sealand. The city of Copenhagen also grew to such a size that an outside beef and milk supply was needed. Since transportation facilities were poor, cattle were walked across the country from the sections in southern Jutland, Sleswig-Holstein, and the west coast in the Ballum section, and many Angler cows were brought in from Angel in southeastern Sleswig. As these cattle moved across the islands, many were traded and sold to the farmers en route. Many of the young calves born en route were left behind because they were unable to keep pace with the drove. As these animals were milked and developed, the Danish farmers realized that these animals were superior to their own native cattle. Consequently, the Danish farmers were dissatis- fied with their poor native cattle, and thus introduced this superior outside blood into their breeding herds. The Red Danish dairy breed, therefore, had its development from 1845 to 1885 by breeding the native cattle to three closely related breeds-~the Ballum, North Sleswig, and Angler cattle. The Ballum cattle came from along the west coast of Sles- wig on the rich grass marshlands of that section. They were char- acterized as a dual-purpose breed, with the weight for a full—grown cow being 1,000 to 1,200 pounds. The color was nearly always a dark red. The Angler cattle are native to the peninsula of Angel, which juts out into the Baltic Sea from the east coast of Sleswig- Holstein. The Angler cattle were a rather small, delicate breed, and milky. The weight of a full—grown cow was from 550 to 775 pounds. Their general color was a light red or medium red, fre- quently with white ‘markings on the belly. A red spotted color was sometimes found. The North Sleswig breed originated in an area lying between, and somewhat north of, the areas of the Ballum and Angler breeds. The size of the breed was a blending of the Ballum and Angler' breeds, with the Ballum cattle having the greatest influence. The Sleswig retained, in part, the milking qualities of the Angler, and yet had a tendency to be beefy, similar to the Ballum. The color, in general, was dark red with nearly black heads and legs. The first recognition of the breed as the Red Danish occurred at the Fourteenth Farmers Assembly (dairy show), Svendborg, Funen, 1878, where, for the first time, there appeared an entry class "Red Danish Cattle of Pure Race.” Prior to that time, the cattle were either called Angler or ”Red. Funen Milk Cattle." The latter of the two early names was rather popular, and actually did not recede until the first Provincial Herd Book was published in 1885 on Sea- land, when the name ”Red Danish Milkrace" was used as the offi- cial designation of all red cattle, which, in a few years, was quite generally accepted all over Denmark. A special impetus was given to the Red Danish breed due to the agricultural crisis of the years around 1880. The Danish grain farmer could not compete with American- and European-grown grains. The price of butter remained high, and a shift was made to dairying. A movement was made to organize cow-testing asso- ciations to help select the best cows and bulls. Consequently, these associations-—the first starting in 1895--spread all over Den- mark and the world. Today Denmark still remains as the leader in cow and progeny testing for herd development. Importation to the United States During the 1930's the United States Department of Agricul- ture conducted an extensive search for plants and animals superior in their production and genetic quality, hoping to introduce them into the United States agricultural system. Since Denmark had the highest average production per cow in the world, importation of the Red Danish Milkrace would be desirable for future study. The av- erage cow in Denmark produced about 100 pounds more butterfat per year than the average cow in the United States. The cows en- tered in the Danish herd book during 1931 had produced an average of 559 pounds of butterfat, some of them being milked three times a day (Reed and Fohrman, 1946). In 1935, Dean E. L. Anthony of Michigan State College was appointed by the USDA to go to Den- mark to select animals for a foundation herd to be used in the United States for study of their production characteristics and trans- mitting qualities. An importation of two bulls and twenty cows was made. The animals came from the Islands Funen and Zealand. The cattle from Funen were selected from two herds in the southern part, While those from Zealand were from two farms in the central part and one in the southernmost tip (Anthony, 1952). This herd eventually was located at the Beltsville Research Center, and has been supervised by specialists of the Dairy Bureau of the USDA. A program was instituted to study the genetic worth of the Red Danish Milkrace . Entrance Into Michigan By 1939, there was a surplus of young bulls produced from this herd. In cooperative arrangement with the Dairy Bureau In— dustry of the USDA and the Extension Service of Michigan State College, bulls from the Beltsville herd were placed with several cooperating dairy farmers in Michigan. According to Baltzer (1952), the project was for the purpose of proving these Red Danish Milkrace bulls in order to make available young proved sires for use in the experimental herd at Beltsville. The program provided that: 1. All cooperating herds would be under production testing in the DHIA. 2. The entire herd would be subscribed to the Red Danish sire project. 3. All females would be kept for at least one lactation record. 4. Replacement of original foundation cows would be made as rapidly as possible with the cross females. 5. All bull calves produced through the project would be castrated. 6. The herds would maintain a Bangs and T.B. health program. Development of the Red Danish Cattle in Michigan On January 25, 1939, a meeting was held in Sanilac County, Michigan, by ten local dairymen to form a Red Dane bull associa- tion for introduction of the Red Dane bulls to be used in the proj— ect. These dairymen subscribed their herds, consisting of 248 cows of the various dairy breeds, to the program. Four of the farmers were appointed bullkeepers for this bull association. On February 17, 1939, four Red Danish bulls arrived at Marlette, Sanilac County, Michigan, to mark the first entrance of Red Danish cattle into Michigan Interest in this project grew among dairymen in other sec- tions of the state. In June, 1939, a Red Dane bull association was formed in Alcona County, Michigan, with twelve breeders subscribing 10 their herds consisting of 213 cows. Oscoda County dairy farmers formed a Red Dane bull association of four bull blocks consisting of thirty herds and 230 cows, in November, 1941. With the advent of artificial insemination, numerous dairy- men professed interest in securing artificial service from Red Danish bulls. In the spring of 1945, Red Dane bulls were placed in artificial service. During the period 1945 to 1948, two hundred ten breeders subscribed 2,233 cows of the various dairy breeds to be bred artificially to Red Dane bulls. The dairymen that sub- scribed their herds to the artificial service agreed to abide by the rules described earlier, just as did the breeders using Red Dane bulls by natural service. In July, 1945, twenty-nine breeders from Ogemaw County, subscribing 386 cows to artificial Red Dane ser- vice, formed a Red Dane Association; followed by thirty breeders in June, 1946, from Calhoun County, subscribing 285 cows; twenty- one breeders in June, 1946, from Isabella County, subscribing 230 cows; forty-nine breeders in July, 1946, from Iosco County, signing 372 cows; twenty-five breeders in October, 1946, from Tuscola County, entering 250 cows; fourteen breeders in December, 1946, from Barry County, subscribing 140 cows; twenty—six breeders in February, 1948, entering 370 cows; twelve breeders in April, 1948, 11 from Newaygo County entered 200 cows; and a natural-service group from Lapeer County entered in 1946. Along with this in- crease of breeders from the above—mentioned counties, the number of breeders in the natural-service blocks also increased. Of the first 2,956 foundation cows entered in this program, 32.8 percent were Guernsey, 29.0 percent were Idolstein, 25.6 per— cent were Shorthorn kept for dairy purposes, 9.2 percent were Jer- sey, 2.7 percent were Brown Swiss, 0.10 percent were Ayrshire, and 0.60 percent were of mixed ancestry. The breed of these cows was identified by the DHIA supervisor. If a cow exhibited enough breed characteristics of any one breed, it was classified as such. All of these cows were grade animals. Origin of the American Red Danish Cattle Association As early as January, 1943, local Red Dane county associa- tion meetings were held and members discussed the possibility of eventual registration of their cattle. In September, 1943, a joint meeting of the Red Dane county associations from Alcona, Oscoda, and Sanilac Counties met to develop standard practices and uniform rules of conduct, and also to discuss and make recommendations for eventual registration. Many meetings were held, and finally 11 from Newaygo County entered 200 cows; and a natural-service group from Lapeer County entered in 1946. Along with this in- crease of breeders from the above-mentioned counties, the number of breeders in the natural-service blocks also increased. Of the first 2,956 foundation cows entered in this program, 32.8 percent were Guernsey, 29.0 percent were‘Holstein, 25.6 per- cent were Shorthorn kept for dairy purposes, 9.2 percent were Jer- sey, 2.7 percent were Brown Swiss, 0.10 percent were Ayrshire, and 0.60 percent were of mixed ancestry. The breed of these cows was identified by the DHIA supervisor. If a cow exhibited enough breed characteristics of any one breed, it was classified as such. All of these cows were grade animals. Origin of the American Red Danish Cattle Association As early as January, 1943, local Red Dane county associa- tion meetings were held and members discussed the possibility of eventual registration of their cattle. In September, 1943, a joint meeting of the Red Dane county associations from Alcona, Oscoda, and Sanilac Counties met to develop standard practices and uniform rules of conduct, and also to discuss and make recommendations for eventual registration. Many meetings were held, and finally 12 on January 16, 1948, the American Red Danish Cattle Association was formed. The purpose of this breed association is to promote the breed of Red Danish cattle and to keep proper records of an- cestry, production of foundation cows, and of the Red Danish descen- dants. An open herd book is to be maintained, providing for the registration of third-cross females and fourth—cross males, provided they meet the requirements set forth by the organization (see Ap- pendix, Exhibit A). GENERAL OBJECTIVES The objectives of this investigation are: 1. To study the effect of three successive crosses of Purebred Red Danish sires on the average milk and butterfat pro- duction, butterfat percentage, and the repeatability and heritability of these characters. 2. To study the contributions of the various Red Danish sires on milk, butterfat, and butterfat percentage based on progeny performances and dam-daughter comparisons. 3. To determine the effect of mild inbreeding as practiced in this population on the yield of milk and butterfat and the occur— rence of lethal defects. 4. To study the milk and butterfat records for variation due to major environmental factors peculiar to this particular popu- lation. GENERAL INVESTIGATIONAL PROCEDURE Source of Data Since it was compulsory that all cows entered in this program of grading up to Red Danish bulls be DHIA tested, the herd books of the cooperating farmers contain, on all cows, all production rec- ords made under this program. In addition to herd testing, an an— nual calving report was submitted to A. C. Baltzer, Extension Dairyman of Michigan State College, in charge of the project. This report contained the ear tag number of the calf, and the number of the calf's sire. The report included all calves born from Septem- ber 1 to August 31 of the succeeding year. ‘ With the incorporation of the American Red Dane Cattle As- sOciation in 1948 by the breeders in this project, a survey was made in Which each breeder listed essential data that was compiled into reCord form by the association. The files of the association con- taimed all data necessary to identify each animal entered in the Program, birth date, sire, dam, progeny and owner. Production data. have been kept by the association since 1949. 15 Colle ction of Data For compilation of data relative to this study, a worksheet designed to include all the desired information for each subject was designed (Fig. l). The American Red Danish Cattle Association cooperated in providing the following data: the breed or generation of the subject, coded as follows: 01, Guernsey; 02, Holstein; 03, Milking Shorthorn; 04, Jersey; 05, Brown Swiss; O6, Ayrshire; 07, Hereford; O8, Angus; 09, Red Poll; 10, mixed breed; 11, first cross Red Dane; 12, second cross Red Dane; 13, third cross Red Dane; 14, fourth cross Red Dane; 15, fifth cross Red Dane; and 16, sixth cross Red Dane; and the eartag number of the subject, date of birth, color by code: 1, red; 2, black; 3, red and white; 4, black and white; and 5, roan; name of herd owner, tattoo number of the sire, and eartag number of the dam. As the Red Dane office completed groups of two. thousand in- C11V'idual cow worksheets, they were returned in numerical order ac— cording to eartag number to the investigator for additional process- ing, Each ear tag number was entered into a herd book and as- signed a code number, the first code number being number 0001, and the succeeding ones being assigned ascending numbers until all were Figure 1 Cow Worksheet Used in Collecting the Data For This Study 55 28 he es s: an elegant 56 no 2 o\ e 661 UGH.39 l8 coded. These code numbers were then entered in the appropriate space on each cow worksheet. The worksheets were then sorted into herd-owner groups. An alphabetical list of the herd owners was made and each was assigned a code number. This number was placed in the respective space on the worksheet. Each herd group was then processed individually for these following data. The worksheets were sorted into ascending year of birth which placed all animals born the same year into the same group. The birth reports for that herd were used to identify the generation and color of each animal which the Red Dane office was unable to supply on about 35 and 95 per- cent of the worksheets, respectively. The groups for each herd were then sorted by the number of the sire. Consequently, all worksheets having the same sire number came into the same group. A code number for the sire was then entered in the appropriate blanks for each sire group. A list of Red Danish sires used in this project had been made previously by the investigator and coded. To assign the same code numbers to the dams as they had when they were S“hiatus, the worksheets were sorted into ascending order by ear- tag numbers, since the code number and eartag number are of di- 1‘ect numerical relationship. With the worksheets in this order, it Was a. relatively easy matter to find rapidly the worksheet of a dam 19 when used as a subject. Then the code number assigned to the ear- tag number as a subject was transferred to the appropriate blank when she became a dam on a worksheet. If a subject in any one particular herd group had a dam that was located in another herd group, the code number was obtained from the herdbook. All dams appeared as subjects within the same herd group at least 95 percent of the time. The worksheets for each herd group were then placed into large manila envelopes along with a set of instructions to help the breeder in completing the desired information (see Appendix, Ex- hibit B). The name and county address of the owner were placed on the outside of each packet. These packets were then grouped together into county groups and sent out to the county agent for each respective county. In some cases the breeders in each county in this program were called in for an instructional meeting of how to properly complete the worksheets; in other instances the county agents visited each breeder and explained the mechanics individually. Fig- ure 1 illustrates the information desired from the breeders. The breeders entered data only beneath the heavy line. Codes were used by the breeders in five instances in completing the worksheet:- disposal of the subject, remarks about each lactation, sex of calf, 20 defects exhibited by calf, and color of calf. Sections 9, 10, ll, 14, and 15 of Figure 2 explain the various codes used for the coding performed by the breeder. The packets of worksheets when completed by the breeders were either returned to the county agents for return to the investi- gator or mailed directly to him by the breeders. As the worksheet packets were returned from the breeders, they were inspected for accuracy and completene s s . Placing of Data on IBM Cards Punching of Data from Worksheets Some of these data from the worksheets were then punched by key—punch operators into IBM punch cards, designated as Indi- vidual Lactation Card No. 1 (Fig. 2), according to the following plan: Columns Punched Data Taken from Worksheet 1—2 Generation Number 3-6 Code Number of Subject 7 Color Code of Subject 11-13 Code Number of Owner of Subject 14-17 Code Number of Sire of Subject 21 Figure 2 IBM Cards 1 and 2 .71 V0 '9N,HSJUJ 90.! .7003 I: 839N/7N 38/9 74 5-78 yo7oa i .1 31/30 / 7273 0V30-2 JAI'IV-I N809 :4 J7VN3J-2 37VN-I X39 ‘3 « k L) < ~I I if 'u \a v) k '5: t L.) Q #/ INDIVIDUAL LACTATION CARD 7VSOdS/0 [SM/:4) 59-59 '99 =55-55 + 29-09 lVJ 'J'N 5'09 X2 29-09 = 55-95 '25 X IS-blr )I7/N 'J'N 5'08 X2 6.9-5.9: Qf-H' X #5589 '29 55-95 '25 UOIJVJ lVJ J0 3397 )I7/N J0 E97 IVO/l V12V7 H1 BM? 7 ~J § 0; re 2*: \r Q2 2 :12 ~12 2’5»: 39V 9NINJHS'JUJ 33 34-3738-39)40 41-43 44-45 49-5/ 52-54 55-59 60-62 63-65p6-6I68-6j7 111/0 9NINJHS‘JUJ ° 939N/7N N0/1 V1.7V7 NVU JO '3 '9'I UJ9IJDN N V0 38/9 JO '3 '9 'I UJ9N/7N JUIS‘ /-/3 1447 [8-20 21-24 25-2I28-29130 UJNMO GHJH '3 '9 'I H 7 409 UJ9I~IDN 1 31/9/75 '9 NOIJ. VUJN39 "‘g LACTAT/ON CARD INDIVIDUAL #2 SUMMARY CARD FRO/‘1 75 '80 NVU JO '3'9'1' 12-52 X #f-Ofi 69- 74 JUIS‘ J0 '3'9'1' 02-9/ XHI-Ot' 62-68 XY OF I. B. C. X MILK PRODUCT/0N 133/‘905 J0 '3'9'1' 0/-8 X ##-0# (uns) 59-59 '99 M) 096*?!» =2>Ir05 + 1.5-5; 62-82 -=‘- 69-98 62-92 -:- 59-06: lVJ 'J'N 50:? X2 (AI/IS) 29-09 M7/N 'J'N 50£x2 (N05) 65-55 28-29 30-35 36- 39 40-44 454748-50 5/‘54 55’6/ (UJQNDN) lNflOJ 08V.) 08V.) 13V7 '0NI NOUJ IZ-L SNND'IOJ UJJSNVUJ 7-27 (mu/m) UJHNIW M03 *7 NOIIVHJNJD 1'": 23 21-24 Code Number of Dam of Subject 28—29 Lactation Number 30-33 Freshening Date—~Year, Month, Day 34—37 Freshening Age--Year, Month 38—39 Calving Interval-—Months 40 Times Milked Daily 41—43 Length of Lactation 44—48 Pounds of Milk 49 51 Pounds of Fat 68-69 Coded Remarks of Each Lactation 70 Sex of Calf 71-72 Defects of Calf 73 Color of Calf 74—77 Sire Number of Calf To facilitate key punching, a mask was improvised which hooded everything but the information desired from the worksheets. A card was punched for each lactation listed for a subject. This included all columns from 1 to 77 as listed in the flow chart. Zeros were punched in columns where information for those respec- tive colurrms were lacking. All cards were punched ”verified” by anothe r operator. Z4 Punching of Inbreeding Coefficients This was accomplished in a rather complicated manner. Without it, however, the task of calculating and entering the coeffi- cients of inbreeding would have been prohibitive for the number in- volved in this study. The inbreeding coefficients of the Red Danish bulls had been calculated and printed on the usual long, folded sheet of paper as will be described subsequently under calculations of inbreeding coef— ficients. A deck of cards—~hereafter referred to as "sire—master inbreeding deck"—-were punched and verified from this list. The code number and inbreeding coefficient of the sire were punched in Columns 14 to 17, and 18 to 20, respectively, one card per sire. The inbreeding coefficients of a subject—-due to the amount of relationship existing between sires--had been calculated as will be described subsequently under calculations of inbreeding coeffi- cients. A deck of cards—-hereafter referred to as ”subject—master inbreeding deck"--were punched from the "subject” worksheets. A card containing generation number in Columns 1 to 2, subject num- ber in Columns 3 to 6, subject's inbreeding coefficient in Columns 8 to 10, subject's sire's number in Columns 14 to 17, and subject's dam's number in Columns 21 to 24 was punched for each subject 25 having an inbreeding coefficient. The "subject-master inbreeding deck” was then sorted behind the ”sire-master inbreeding deck" on Columns 14 to 17. This placed all the cards of daughters of'a particular sire behind his card. The inbreeding coefficients in the ”sire-master inbreeding deck” were transferred to the respective columns in the ”subject-master inbreeding deck"--Columns 18 to 20—-by gang—punching. The inbreeding coefficients of the subjects were partial to the extent of needing to be multiplied by the factor ((1 + FS)(1 + Fd), where FS is the inbreeding coefficient of the sire and Fd is the inbreeding coefficient of the dam, as described under cal— culations of inbreeding coefficients. To accomplish this cor— rection it was necessary to sort the ”subject—master inbreed- ing deck" on Columns 1 to 2. This placed the second, third, and fourth generation cards in respective decks. The second generation was the first generation to have inbreeding coeffi- cients; the correction that needed to be made was the factor 1 + F , 5 since the dams were not inbred. The second generation's ”subject- master inbreeding deck" cards were sorted on Columns 18 to 20-- sire's inbreeding coefficient--behind a deck of cards, hereafter called “sire inbreeding correction deck," containing the factor 26 needed to correct for each degree of inbreeding in Columns 18 to 20. The "sire inbreeding correction deck" had been made from a table of factors developed by the investigator which was, in sub- stance, all combinations of f(l + FS)(1 + Fd" Each "subject- _ master inbreeding deck" card fell behind the “sire inbreeding cor- rection deck" card having the same amount of inbreeding in Columns 18 to 20. The correction factor was then gang-punched in Columns 28 to 30 of the "subject—master inbreeding deck." The "subject- master inbreeding deck" was then processed through the 602 Cal- culator. Columns 8 to 10--containing the subjects' partial inbreed— ing coefficient——were multiplied by Columns 28 to 30-—containing the correction factor, and the sum-~the "whole" inbreeding coefficient --was punched in Columns 31 to 33. ’ The third-generation Red Danish females have as dams the same females appearing as subjects in the second generation. There- fore, it was only a procedure of collation to transfer the inbreeding coefficients from the cards of the second generation‘s "subject- master inbreeding deck" to the columns designated for the dam's inbreeding coefficient, Columns 25 to 27. The second generation's "subject-master inbreeding deck” was ordered on Columns 3 to 6, and the third generation's ”subject-master inbreeding deck" was 27 ordered on Columns 21 to 24. When the subject's number appearing in Columns 3 to 6 of the second-generation deck was the same as the dam number in Columns 21 to 24—«a second-generation subject being a dam in the third generation, the inbreeding coefficient ap- pearing in Columns 30 to 32 of the second-generation deck was transferred to Columns 25 to 27 of the third-generation deck. The set of third-generation "subject-master inbreeding deck" cards con- tained, with respect to inbreeding coefficients, a partial inbreeding coefficient of the subject in Columns 8 to 10--the sire's inbreeding coefficient in Columns 18 to 20 and the dam‘s inbreeding coefficient in Columns 25 to 27. The procedure to adjust the inbreeding coef— ficient of the subject from a partial factor to a whole factor was essentially the same as followed with the second generation's "sub- ject—master inbreeding deck" cards. The procedure differed slightly, however, since in the third generation, it was possible for the dams to have an inbreeding coefficient whereas in the second generation they did not. This necessitated a slightly different correction factor from that obtained in the expression [(1 + Fs)(1 + Fd). A “sire- dam inbreeding correction deck" was punched from the table of correction factors computed by the investigator. The third genera- tion's "subject-master inbreeding deck" was sorted on Columns 18 27 ordered on Columns 21 to 24. When the subject's number appearing in Columns 3 to 6 of the second—generation deck was the same as the dam number in Columns 21 to 24--—a second-generation subject being a dam in the third generation, the inbreeding coefficient ap- pearing in Columns 30 to 32 of the second-generation deck was transferred to Columns 25 to 27 of the third-generation deck. The set of third-generation "subject-master inbreeding deck" cards con- tained, with respect to inbreeding coefficients, a partial inbreeding coefficient of the subject in Columns 8 to 10——the sire’s inbreeding coefficient in Cqumns 18 to 20 and the dam's inbreeding coefficient in Columns 25 to 27. The procedure to adjust the inbreeding coef- ficient of the subject from a partial factor to a whole factor was essentially the same as followed with the second generation's "sub- ject-master inbreeding deck" cards. The procedure differed slightly, however, since in the third generation, it was possible for the dams to have an inbreeding coefficient whereas in the second generation they did not. This necessitated a slightly different correction factor from that obtained in the expression n1 + FS)(1 + Fd). A "sire- dam inbreeding correction deck" was punched from the table of correction factors computed by the investigator. The third genera- tion's "subject-master inbreeding deck" was sorted on Columns 18 28 to 20, behind the ”sire-dam inbreeding correction deck." Then each deck resulting from the sort on Columns 18 to 20 was sorted on Columns 25 to 27. This placed all ”subject-master inbreeding deck” cards behind the "sire-dam inbreeding correction deck" card containing the factor needed for multiplication to correct the partial inbreeding coefficient of the subject. The procedure for ensuing generations was identical with that of the third generation. The subject lactation cards and the ”subject-master inbreeding deck" cards were then ordered on Columns 3 to 6. By collation, the inbreeding coefficients appearing in Columns 14 to 17, 25 to 27, and 30 to 32 of the "subject—master inbreeding deck“ cards were transferred to Columns 14- to 17, 25 to 27, and 8 to 10, respectively, of the subject lactation cards. The subject lactation cards then contained the inbreeding coefficients of the subject, sire and dam in Columns 8 to 10, 14 to 17, and 25 to 27, respectively. Standardization of Records for Age and Length of Records The data utilized here are the production records of milk, butterfat and butterfat percentage, and ages and dates that con- Cemed each record. All records of a cow--of 270 to 365 in 29 1ength--up to and including her tenth lactation were used. The amount of information derived regarding a cow's productive ability rises at a decreasing rate with additional records, particularly after the third record. Records at very advanced ages usually con- tain a considerable amount of error due to permanent environmental changes such as damage to mammary tissue (Berry, 1945). Berry and Lush (1939) and Lush _e_t_ fl. (1941) discussed the question of omission of records from a cow's average and concluded that the omission of a record known to be made under abnormal circum— stances might increase the value of the average. In this study, all records were used, except those followed by abortion 154 days after breeding, or where the record carried notation of an illness or ac— cident deemed serious enough, by the investigator, to affect produc-- tion. Such records were identified by a code number and sorted out with an IBM sorter, so that omission of a record was imper- sonal and not affected by its size of production figures. The life- time average was used as the measure of each cow's producing ability. Berry and Lush (1939) and Berry (1945) proposed a method of correcting for incomplete repeatability where unequal numbers of records are used for the average production of a cow. This method places such averages on a comparable basis and gives JCHLL F, .. 1L1i ‘ 30 greater weight to deviations from the herd average, where the devi— ation is based on larger numbers of records. However, Laben and Herman (1950) used this method, as well as the lifetime average of a cow, for comparing dams and daughters and estimating heritability of milk and fat yield and found no increased correlation when using the corrected lactation numbers formula. After weighing all of the facts and theories, it was deemed best to use the lifetime average production figures in the statistical analysis of these data. Lactation records from 270 to 365 days' length were stand— ardized up or down to a 305—day basis, according to correction factors developed by Dickerson (1941). Corrections for age at freshening were made by application of age correction factors de— veloped by Kendrick (1942). The two factors for age and length of lactation were multiplied together for all possible combinations of age and length of record to secure a single correction factor for easier calculation with the 602 Calculator. Each of these combined factors was then punched in individual master cards, Columns 52 to 54, each master being identified by a breed code number the same as used to indicate breed or generation of the subject. All masters were X'd in Column 74. The mixed breed correction factors were used for all Danish crosses. The following procedure was used to 31 insert the correction factor required to convert each lactation record to a 2X 305-day mature equivalent record into the Individual Lactation Card No. 1 (Fig. 2). 1. All individual lactation cards were sorted on Column 1. All cards falling in the 1 pocket were either mixed—breed foundation cows or Danish crosses. These were retained together, since the same correction factors were used for both groups. The cards falling on the 0 pocket were then sorted on Column 2. This sort separated these cards into breed decks; these decks were then com- bined into groups by the order of Ayrshire, Guernsey, and Jersey; Brown Swiss, Milking Shorthorn and Red Poll; and Holstein, re— spectively, since the correction factors used were computed in that order. This was done by combining all cards falling into the 1, 4, and 6 pockets for the Ayrshire, Jersey, and Guernsey group, com- bining all cards falling into the 3, 5, and 9 pockets for the Brown Swiss, Milking Shorthorn, and Red Poll group, and keeping the cards falling in he 2 pocket separate as the Holstein group. 2. Each of the above decks were sorted separately to fresh- ening age--Columns 34 to 37--and to length of lactation--Columns 41 to 43-.-behind the master cards containing the correction factors for the specific group. This procedure placed all lactation cards 32 needing a specific correction factor for age at freshening and length of lactation behind the master card containing this factor. The cards were then gang-punched, the correction factor transferring from Columns 52 to 54 of the master card to Columns 52 to 54 of the individual lactation cards. 3. The master cards were then separated from all decks by sorting for X in Column 78. 4. The individual lactation cards were then processed through the 602 Calculator by the following calculations: a. Columns 44-48 X 52,-54 into 55-59 = 305 2X M.E. milk production. b. Columns 49-51 X 52.-54 into 60-62 = 305 2X M.E. butterfat production. c. Columns 60-62. é 55—59 into 63-64 = Percentage of butterfat. (The periods indicate decimal points.) Calculation of Average Lifetime Production In order to place all records of a cow on an average lifetime production basis, it was necessary to make another card designated as Summary Card No. 2 (Fig. 2). The data used to make this 33 card were obtained from Card No. 1, the individual lactation card. Each cow had one card. The procedures were as follows: 1. All individual lactation cards were sorted on Columns 3 to 6. 2. The cards were in order by subject number, with all individual lactation cards of a particular subject in sequence. 3. The tabulator was utilized to make a summary card. When the subject number changed a summary card was punched for each subject showing: a. Columns 1—24 = Group transfer of same data from Card No. 1. b. Columns 28--29 : A card count which indicated number of lactations being summarized. c. Columns 30-35 = Total lifetime 305 2X M.E. milk production, a summation of individual lactations from Card No. 1, Columns 55-59. d. Columns 36—39 = Total lifetime 305 2X M.E. but- terfat production, a summation of individual butter- fat yields from Card No. 1, Columns 60-62. 4. The 602 Calculator was then utilized in order to obtain the average lifetime production of milk and butterfat and calculation E 34 of the butterfat percentage from the total 305 2X M.E. production of milk and butterfat found in Columns 30 to 35 and 36 to 39, re— spectively. The calculations were as follows: a. Columns 30—35 4 28—29 into 40—44 = average life-— time milk production. II b. Columns 36—39 {- 28—29 into 45—47 average life- time butterfat production. c. Columns 45—47 4 40—44 into 48.49—50 = average lifetime percent of butterfat. 5. Since Card No. 2 had unused columns, these columns were utilized for the calculation of regression of milk production on inbreeding. These calculations gave the cross-product terms for inbreeding X milk production and were as follows, when X = inbreeding coefficients, Y = production: a. Columns 40—44 X 8—9.10 into 55—61 = XY of sub— ject. b. Columns 40—44 X 18—19.20 into 62-68 2 XY of sire. c. Columns 40—44 X 25—26.27 into 69—74 = XY of dam. 35 Calculation of Total Sums of Squares It was desirous to calculate the total sums of squares for milk and butterfat production, since these statistics are required for calculation of the analysis of variance, intraclass correlation, standard deviation, ”t" test, and analysis of covariance using intrasire dam—daughter regression, dam—daughter regressions, dam—daughter correlations, paternal half—sib correlations, and re— gression of production on inbreeding. A card, Repeatability Total Sum of Squares Card No. 3, was made in order to perform the calculations (Fig. 3). Each cow had as many cards as she had lactations. The procedures were as follows: 1. All data used in this card were transferred from Card No. 1 in the following sequence: a. Columns 1—2 into 1—2 Generation of subject. Code number of subject. b. Columns 3—6 into 3—6 He rd owner numbe r. H c. Columns 11—13 into 7-9 Sire number of subject. (1. Columns 14—17 into 10—13 e. Columns 21-24 into 14—17 Dam number of subject. 305 day 2x M.E. milk f. Columns 55—59 into 18—22 p roduction. 36 Figure 3 IBM Cards 3 and 4 56-80 68-98 “IVA 83.1 NI 9N/A 7 VJ £8 '08 .71 V0 9N/NJHS‘JUJ 295-5; '55 21w 111/33de 9f '5} if OlNI 59-159 99 1V.-/ 1NJJUJd P— elf-.99 leJ 'S‘,97 ZE'SE OlN/ 29-09 .LVJ 'J'N 506‘ X? 35-37 38 43 171‘: 77 49 50-53 54-5 222-9] 2 M7IN 2297 22-9/ 01N/ 69-529 )l7/N '3'N308 X? Zl-f/ OiNl #Z-IZ UJQNDN NVU €l-0/ OiN/ U-H HIGH/7N 38/3 6'1 OlNl €l-II HEN/‘10 083/1 9-5‘ OJN/ 9'8 UJGNIIN M03 Z-I OJNI Z-I NOI1Vt/3N39 ['2 3'6 7'9 IO'I3 H'l7 [8'22 23’34 #4 REPEATAB/L/TY BETWEEN 22(TRAN5FERRED FRO/‘1 CARD #2 09 '61-21 =9/-.9/ + 91-IL 77-60 91-54114; al/w zoz-osss-zs 02-69'99-[9 0.1.N/ 5.9-8.9 'ZS-IS 99-59 0.l.Nl El-II UJNMO OUJH 64'6'6‘6 7'70 7/ ' 76 9/-$/ -Z— 95-6fi .LVJZK N33M139 57-63 2-95 -55 31w 597? 71/101 49'56 95-55 0.LNI 68-98 (7V101)1VJ'3'N .505 x2 45' 18 9/-5/ + 55-52 mm 2:: NJJM139 35—44 aza-z/ 2M7IN 5,97 275201 23'34 224/ 01N/ 55-0? (717101) X7!“ '3'” 502 X? 9/-$/ OJNI 62-9? (HIGH/7N) anoa MOD fl-II 01M] 53']? UJHNDN NVG OI-L 01N/ ll-H UJUNHN 38/9 HG ll I4 I: {5 [7°22 9-E 01M! 9'8 UJQNHN M09 2'! 01M] 2-/ NOIl VUJN39 I'Z 3'5 data: Th CD 38 Columns 60~62 into 35—37 = 305 day 2X M.E. but— terfat production. Columns 63.64, 79 into 44.45, 46 = Percent of but— terfat. Columns 30—33 into 50—53 = Date of freshening. Columns 38—39 into 54—55 2 Length of previous calving interval. following 602 calculations were performed with the 2 Columns 18—22 into 23—34 = X2 for milk production. 2 , 2 Columns 35—37 mto 38—43 = X for butterfat pro— duction. Z 2 Columns 44.45, 46 into 47—4849 = X for percent— age of butterfat. Calculation of the Between Sums of Squares The between sums of squares were necessary for each of alyses used to analyze these data. Therefore, these statistics :alculated by making another calculation card, Repeatability :n Sum of Squares Card No. 4 (Fig. 3). This card served iish all between sums of squares needed and not just the 39 lility analysis—-as the name implies. This was merely an ation of the card for working purposes. Bach cow had one card and the procedures necessary for ulation of this card were as follows: All data used in this card were transferred from Card L the following order: lata: a. b. Columns 1-2 into 1— Columns 3—6 into 3— Columns 14—-17 into Columns 21—24 into Columns 28—29 into lactations) . Columns 30—35 into milk production. Columns 36-39 into production. Columns 11—13 into 2 Gene ration of subject. 6 = Code number of subject. 7—10 11—14 15—16 17-22 45—48 64—66 Sire number of subject. N Dam number of subject. Card count (number of Total 305 2X ME. Total 305 2X M.E. fat Herd owner of subject. Columns 51—52.53—54 into 67, 68.69—70 = Total percent butte rfat. The following 602 calculations were performed with the 4O ' 2 z a. Columns 17—22 into 23—34 = S(X) for milk pro— duction. b. Columns 23—34 —2— 15—16 into 35-44 = [S(X)z]/n for milk production. 2 2 c. Columns 45—48 into 49—56 = S(X) for butterfat production. (1. Columns 49-56 -'.- 1516 into 57—64 = resulted in [S(X)Z]/N for butterfat production. ~ 2. e. Columns 67, 68.—69, 70 into 71, 74 .75, 76 == S(X)z for butterfat percentage. f. Columns 71—76 -2- 15—16 into 77—78—80 2 [S(X)2]/n for butte rfat pe rcentage . Calculation of Regression of Production on Inbreeding The data needed for this analysis were calculated on a card ed as Regression of Production in Inbreeding Card No. 5 ). The data needed for the calculations were transferred he Summary Card No. 2, since average lifetime production 5 were used. One part of the statistic was calculated on lo. 2, as calculating space was not available on Card No. 5. 41 Figure 4 IBM Cards 5 and 6 O UJONDN 3915 ”P 25-95 x 51-51 0 ‘52 ilk/755% x '50 v0 1’ 1 .10 . 252-12 1? "x " 2 0061a! .LVJ ’3/W g t, 01-59 011w 51-59 g a OSJMZVJ 14'» 30 d 202-9! 053 a b 2.00Ud ”WI-J 'JAV 6 5 ‘° > 59-95 01m 99-09 g E g (GAV) I Q g 3' 3‘ g'GOUd X7/N :3 92-12 [51-51 u; g ‘ C (LIVU) lVJ X 3 HI :3- (,3 y) ““5941; 0./.N/ [9-5; :3 . . t #JAV 5,11v0 '0099 NJ - 59 5. ‘° ' Q “'35 92-12 x 21-01 3; t 35 25-95 011w 55-05 a; L. (32115) 1w x 091 as \ U .3” “JV” «'0 V \ h. o >‘ m E 15-95 011w 52-92 3? xx 52-12 xe-z. 1; IE ””03 0W? ‘3 (re/Iglw x .991 g. CT: “4.5 x 2&0? ,2 g Lu suva X'snva . \ It 1w .10 AX :3 aSI-El Y- gNVU .10 9'91 :9, In 15-29 011w 51-59 v- w Q a ('JAVI ‘P 32/-01 8 2 Q aGOUd ”.1 % 3.71115 .10 0'91 .9, < i} 9 12-92 011w 99-09 - o- . "5 36“ 3' 5 'aozij/‘y'uu ‘i’ guy/‘90: .10 '09 I :5 c2: 75 E 2 N ‘I 52-12 011w 15-55 "-9 \ q 22-02 01m 15-55 m 009d 1119 31w E hp, '3Av mm '1‘ m . - O 0 02-91 011w 55-05 8 § 3; a W "J N I ”03" ”7’“ 34V ‘3 an E “‘ 51-51 0111/ 55-05 a. 51-51 011w 12-52 “9 <( K9 it 3115' 50:70 T mm .10 '0 '9 'I e > 3Q °0oad 117151 '2 21-01 011w 02-91 § “-1 o 0: 51-11 011w 52-12 a 32115 .10 '09 'I 2 5 assumv mm : 5-1 011w 01-9 3- Lu 01-1 011w 11-51 2 g 1331‘9/15‘ .10 9'91 v- Q: L- ussumv 3915 :L 9-9 0111/ 9-9 <9 «7, 3’: 51-: 01M 9-5 «3 UJUNHN 133F903 5') ‘0 k UJQNDN M03 to 2-1 0111/ 2-1 3., 2-/ arm 2-1 N . N0 11 VUJNJQ ~ 11: NULL van/v.79 L 43 The procedure of card make—up and calculations was as 1. All data used in this card were transferred from Sum-— Sard No. 2 in the following order: a. Columns 1-2 into 1-2 Generation of subject. b. Columns 3—6 into 3-6 = Code number of subject. c. Colunms 8—10 into 7—9 = Inbreeding coefficient of subject. Inb reeding c oefficient (1. Columns 18—20 into 10—12 of sire. e. Columns 25-27 into 13—15 Inbreeding coefficient of dam. f. Columns 40—44 into 16—20 Average lifetime milk production. g. Columns 45—47 into 21—23 Average lifetime butter— fat production. 11 h. Columns 14-17 into 75—78 Sire's number. 2. The following calculations were made with the 602 Cal-- to obtain the X2, Y2, and the cross—product terms of in- g times butterfat production, where X = inbreeding, Y = ion. 44 a. Columns 7—9'Z into 24-29 = X2 of subject. ' b. Columns lO--122 into 30-35 2 c. Columns 13—15 into 36-—4l 2 X of sire. X2 of dam. (1. Columns 7—9 X 21—24 into 42—46.47 = XY of subject. e. Columns 10—12 X 21-23 into 48—51.52 f. Columns 13—15 X 21—23 into 53—56.57 2 g. Columns 16—20 into 58—66 production. 2 h. Columns 21—23 into 67—71 te rfat p roduction. XY of sire. XY of dam. Y2 of subject's milk 2 Y of subject's but- Calculations of Sire Evaluation and Estimates of Heritability The data involved in this card, the Sire Evaluation and ability Card No. 6 (Fig. 4), were transferred from two cards, No. 2 and Card No. 5. 1. The following data were transferred: a. Columns 1—2 into 1—2 b. Columns 3—6 into 3—6 c. Columns 14—17 into 7— Gene ration of subject. Code number of subject. 10 = Sire’s number. d. Columns 21—24 into 11—14 = Dam's number. 45 e. Columns 40—44 into 15—19 Subject's average life— time milk production. Subject's average life— f. Columns 45—47 into 20—21 time butte rfat production. Number of lactations H g. Columns 28—29 into 46—47 of subject. 2. The cards from Step 1 were then sorted to Columns 11 which placed them in order by dam number. They were then :1 against Card No. 2, which was sorted to Columns 3 to 6. atched a subject of Card No. 2 with her daughter's Card No. : following information from Card No. 2 was then transferred natching Card No. 5 (Columns 3 to 6 of Card No. 2 matching 3 11 to 14 of Card No. 6): a. Columns 40—44 into 48—52 Dam ' 3 average lifetime milk production. Dam ‘ s ave rage lifetime b. Columns 45—47 into 53—55 butterfat production. 3. The cards were then separated and Card No. 6 from 1nd 2 was sorted to Columns 3 to 6 and matched with Card orted to Columns 3 to 6. The following information was ’red from Card No. 5 to its matching Card No. 6 (Columns natching 3 to 6): 46 a. Columns 57-65 into 23—31 Subject's average milk p roduction s qua red. b. Columns 67—71 into 32—37 Subject‘s average but— terfat production squared. 4. The cards were then separated and Card No. 6 from s l, 2, and 3 was sorted to Columns 11 to 14, which placed 1 in order by dam number. Card No. 5 was sorted to Col—- 3 3 to 6 and the two cards matched. The following data were sferred from Card No. 5 to its matching Card No. 6 (Columns 0 14 matching 3 to 6): \ a. Columns 58—66 into 56—64 Dam's average milk pro— duction squared. b. Columns 67—71 into 65—70 Dam's average butterfat production squared. 5. Card No. 6 from Steps 1, 2, 3, and 5 was then processed igh the 602 Calculator for the following, where X equals dam's uction and Y equals daughter's production: a. Columns 20-22 X 53—55 into 38—45 XY of daughter- dam milk production. 11 b. Columns 15—19 X 48—52 into 71—80 XY of daughter—- dam butterfat production. 47 Methods Used in Deriving Coefficients of Inbreeding Bulls A six—generation pedigree was constructed for all Red Danish is used in the project. This necessitated the coding of many ani— ls since the pedigrees of the foundation stock were filled in with nes only and it was desirous to have numbers for IBM calcula— IS. The measure used to calculate the intensity of inbreeding the one developed by Wright (1922) as follows: Fx == Z[(1/2)n j- + 1(1 + Fa)] where Fx is the inbreeding coefficient of animal x; .nd n' are the number of generations from sire and dam, respec— 2ly, to the common ancestor, and Fa is the inbreeding of the estor itself. However, in the calculations of these data, a ghtly different written formula was used, although the numerical ults were identical with those of Wright's original formula. As ised for IBM technique by Hazel and Lush (1950), the formula is re conveniently written here as Fx = 22((1/2)n + n'(1 + Fa)] :re n and n' are defined as the numbers of generations from sub— t x through its sire and dam, respectively, to a common ancestor. 2 change in definition of n and n' is, in this case, solely a 48 nvenience which permits the same cards to be used for computing :her the inbreeding of x or the relationship of x to some other imal. As previously stated, the coefficients of inbreeding were cal— lated for these data by the IBM procedure described in detail by Lzel and Lush (1950). However, this procedure was improved in .e instance, and a further procedure was added to eliminate all .ndwork. This last operation by machine reduces chances of er- ms in at least five instances——reading, posting, copying, factoring, Ld reposting——when large numbers are involved, such as finding e coefficients of inbreeding and relationships of eighty-seven ani— als as in these data, with 14,847 paths of relationship. The procedure as outlined by Hazel and Lush reads in one stance as follows: "The cards are machine sorted into hereditary coups ordered numerically by ancestor number, or alphabetically ,r ancestor name if ancestors are not numbered. The subgroups ider each ancestor are ordered by ancestor's offspring. Within ich subgroup the cards are in order by subject number.” This rocedure was changed in the calculations of these data in the fol— »wing ways. The cards were machine sorted in inverse order into ereditary groups ordered numerically by subject number, in this 75 i 1.‘1Li 49 3e no names being used. The subgroups under each subject were iered by ancestor's offspring. Within each subgroup the cards re ordered by ancestor's number. In Hazel's and Lush‘s procedure a two—way table is set up and : n + n' values are written in the table as they occur on the hered— list. After completing this table, the coefficients for each com— iation of subjects are reduced by combining powers of one—half to :ir smallest terms and are then converted to inbreeding coeffi— :nts and genic covariances and posted in another two—way table. In calculating the inbreeding coefficients and genic covariance r these data it was decided that to work with two 87 by 87 two— .y tables and posting and combining approximately 15,000 n + n' lues would resolve into too many errors and be beyond the scope this study in regards to clerical work involved. Therefore it .s decided to formulate a procedure to eliminate these two tables, d, as an end result, have the inbreeding coefficients and genic variances tabulated on the usual long, folded sheet of paper. The :thod evolved consisted of the following steps. 1. Using the hereditary list, a card was punched for each bject and its related subject showing the generation numbers + n') of common ancestor to subject and related subject. By 50 other method these generation numbers (n + n') would have en posted in a two-way table in the cell where the paths of the bject and related subject cross. When punching these relationship rds, the smaller subject number should be kept in the left-hand :ld. An example is shown in Figure 5. The sample card illus— ated in the upper section of Figure 5 is taken from the hereditary oup listed in the lower section. All subsequent cards were punched the same manner. 2. There are two possible methods for punching the relation- Lip cards: key—punching and selective reproducing and gang—punch- g. Due to the wide variations of the number of listings in the Lrious hereditary groups, the most efficient method can only be :termined after a preliminary check of the listings of the hered— ary groups. In rather extensive hereditary groups and with numerous icestor offspring subgroups within an ancestor group, the best Lethod would probably be that of key—~punching. Although this is Le least desirable method of the two, it does, however, cancel four 1 the five chances of errors possible in using the two—way tables. ’ the cards are punch—verified, the chances of reading error is [so eliminated. 51 Figure 5 Example of Relationship Card Columns 1-4 5—8 9 10 Generations from Generations from , Related , . .bject , Subject to Common Related Subjects to Subject Ancestor Common Ancestor 1504 1505 2 5 A t ' S'd bject nces or S Ancestor 1 f3 Of Generation Offspring Pedigree 504* 2000 1001* 1 2 505 2000 1001 1 5 504* 2023 1001 l 4 505 2023 1001 2 5 507 2021 1001 1 3 509 2021 1001 2 5 * No relationship cards were punched for a subject re- ed to itself when the common ancestor was on the same side the pedigree. The above example in the hereditary list illus- .tes this in the case of matching 1504 with 1504 because the nmon ancestor is on the same side--side 1 or paternal side. e side does not- enter into the matching of different subject num- '5. ,e remains as 1. For example, 1504 would match with 1507 even though the This means they are related through the top .8 of the pedigree. .ili 52 The other method used in this study consisted of taking the riginal deck of hereditary cards, sorting the group to ascending rder on the columns containing the subject number (1 to 4 for lustration), and then working from the hereditary group list. The rst card that appears on the list was taken as a master. The ards to be matched with this card were divided into two groups-— 108% which contain subject numbers smaller than the master and rose which contain subject numbers greater than the master card. he group with the smaller subject numbers were reproduced into ew cards-—Columns 1 to 4 to Columns 1 to 4--also reproducing 1e generation number. Then the subject number from the master ard was gang—punched into the new cards in the appropriate col— mns. The generation number was also gang-punched from the aaster card. The group of cards with the subject numbers larger ban the master card was processed the same way with the exception if reversing the fields in the new cards so that the smaller subject Lumber of the master card remained on the left. 3. After the relationship deck had been completed, the cards vere sorted into decks containing the same total of the two genera- :ion numbers appearing in each card. This was done by the fol— .owing procedure. The cards were sorted on one of the columns 53 containing a generation number. The sort results in six decks, 0 through 5 when working with 5 generation pedigrees. These decks were kept separate and each deck was sorted on the other column containing the second generation number. All decks except the 0 deck broke down into 6 decks, the 0 deck into 5. The cards of the second sort were handled in the following manner: a. Since the column sorted contained only numbers from 0 to 5 (because of using 5 generation pedigrees in this project), at each sort cards fell only into the O to 5 pockets. b. The cards from the 0 deck sort were dropped in storage pockets .1, 2, 3, 4, and 5; cards from the 1 deck sort were dropped in storage pockets 1, 2, 3, 4, 5, and 6; the cards from the 2 deck sort were dropped in storage pockets 2, 3, 4, 5, 6, and 7; cards from the 3 deck sort were dropped in storage pock— ets 3, 4, 5, 6, 7, and 8; cards from the 4 deck sort were dropped in storage pockets 4, 5, 6, 7, 8, and 9; and the cards from the 5 deck sort were dropped in pockets 5, 6, 7, 8, 9, and 10. 54 c. All cards in storage pocket 1 had a generation total of (1/2)]; cards in storage pocket 2 had a generation total of (1/2)2 and so on, for cards in pockets 3, 4, , 7, 8, 9, and 10. 4. A master card with a coefficient corresponding to the generation total was placed with each stack of cards from the stor— age pockets and this coefficient was gang—punched into the selected columns. 5. The relationship cards were then sorted to ascending order by sorting on subject number and related subject number using the .two as one number. 6. This sort was then tabulated with minor control on sub— ject relation number. The printed list will have the coefficient of relationships of any one animal with all other animals in the study and the coefficient of relationship to itself. An example is given in Figure 6. Females The measure most commonly used to calculate inbreeding coefficients has been referred to on pages 47 and 48. Lush (1949) presented a method whereby the genic covariance between sire and 55 Figure 6 Relationship Li st Subject Related Subject Relationship 1507 1507 . 1250* 1507 1508 2650 1507 1509 3356 '1507 1510 5625 1508 1508 0625* 1508 1509 1462 1508 1510 1325 1508 1511 2121 1508 1512 1913 * This relationship of an animal to itself multiplied by 2 will give the correct F for any particular animal. The proof for this was given on pages 47 and 48. 56 dam of subject X is used. In this method Fx = [de/Z] X /(1 + FS)(1 + Fd), where Fx is the same as in Wright's equation, de is the relationship between sire and dam, and /(l + F8)(1 + Fd) is the square root of one plus the sire's inbreeding coefficient times one plus the dam's inbreeding coefficient. Fx may likewise be determined by the summation of certain fractions of the genic covariance between the sire and all of the sires appearing in the bottom half of the pedigree. The equation for this method may be conveniently written as Fx = RsMgsi/4 + RsMgsj/8 + RsMgsk/16, depending on number of generation used, times fil + FS)(1 + F ), d where RsMgsi/4 is the one—fourth of the genic covariance of sire and maternal grandsire, RsMgsj/8 is one-eighth of the genic co— variance of sire and maternal great grandsire, RsMgsk/l6 is one- sixteenth the genie covariance of sire and maternal great, great, grandsire, and /(l + FS)(1 + Fd) is the same as defined earlier. Since the genic covariance of all Red Danish bulls used in this project had been calculated and printed on the usual long, folded sheet of paper, it was decided to use the last described method for the calculation of the inbreeding coefficients of the fe— male crosses. 57 The females in this study were the progeny of a succession of four generations of Red Danish sires bred to the various dairy breeds and to the subsequent resulting cross females. Therefore the first generation females were not inbred due to the out—cross of Red Dane to an unrelated breed. The second generation females, however, were inbred if the sire and dam's sire, both Red Danish, were related. The calculation equation for the inbreeding of this cross can be written as Fx = RsMgsi/4 X /(l + Fs)(l + Fd). In the third generation, there were two Red Danish sires in the bot- tom half of the pedigree, consequently the equation for calculating these inbreeding coefficients was Fx = [RsMgsi/4 + RsMgsj/8] X /(l + F )(l + F ). Since in the fourth generation female pedigrees, s (1 three Red Danish sires appeared in the bottom half of the pedigree, the equation for the inbreeding coefficients was Fx 2 [RsMgsi/4 + RsMgsj/8] x [(1 .+ FS)(1 + Fd). The part of the equation due to the genic covariance of the sires in question was obtained from the genic covariance listing of Red Danish sire, previously explained. It was a matter of deter—- mining the sire and the sires on maternal side of pedigree for the various crosses and by referring to the genic covariance listing and dividing the product listed as the relationship between the sires in 58 question by the proper divisor. These relationships could also be multiplied by the proper fraction, 0.25, 0.125, or 0.0625, as the case may be, for easier calculation. For the second, third, and fourth generation animals, it was conveniently done by using the ac— cumulation multiplier and calculating a final product which was the part of the inbreeding coefficient due to the relationship between sire and dam. Unless the sire and the dam are highly inbred, the term under the square root sign will not differ much from 1.0; it would be ap— proximately correct to say that the inbreeding of the offspring is one—half of the relationship between its parents. However, in this study the term under the square root sign was calculated and multi— plied times the term due to the relationship between the parents. To facilitate the addition of the part of the inbreeding coefficient due to the term under the square root sign, a table of calculated values was constructed and a "sire and dam correction inbreeding deck" of IBM cards was punched from the table to enable addition of the term by IBM processes. This is described under punching 0f inbreeding coefficients. AVERAGE IVIILK AND BU'I‘TERFAT PRODUCTION, REPEi T— ABILITY, AND HERITABILITY OT? COWS RESULTING FROIxI THREE SUCCESSIVE CROSSES Oi? PUREBRED RED DANISH SIRES Review of Lite ratu re Grading—Up Effect Introduction Observations by breeders have long established that cross— breeding corrected the deleterious results often found in inbreeding, in which the undesirable recessive genes come together in the homo- zygous phase. Observations have also indicated that mating animals less closely related than that occurring in random breeding produces offspring with increased size and production as compared to inbred or purebred animals. In 1.914 it was suggested that this increase be called heterosis (Shull, 1948). The American Society of Animal Production Committee on Investigations (1940) accepted heterosis as being synonymous with hybrid vigor from the standpoint of its usage in animal breeding. They define heterosis as the superiority that is exhibited by the progeny over the better parent. This applies 60 to the progeny from crossing of strains, lines, varieties, breeds, and species. A brief, recent review of the different degrees of crossbreed— ing animals is given by Gilmore (1952). In general these experi— ments with subgenera, species and breed~crossings indicate that higher production can be obtained in fewer generations than with selection within the breed. However, the results to date indicate that just as great performance can be obtained with continued se— lection within the breed. A longer time is required, however, and culling must be judicious and continuous. Reed (1951) reported that the first ten Red Sindhi—Jersey heifers tested at Beltsville aver— aged 8,717 pounds of milk and 515.6 pounds of butterfat actual pro- duction milked three times daily, mostly 365 days in length. A total of 139 crossbred females of various ages were on hand at the end of June, 1951. Of the total number, one hundred fourteen are 50 percent Sindhi, fifteen are 75 percent Sindhi, and ten are 25 percent Sindhi, the rest of their breeding being Jersey, Brown Swiss, or Holstein. The ultimate objective of that breeding project is to develop heat—resistant strains of milking cows for the south. The purpose of this study was to determine the effect of three successive crosses of Red Danish bulls on average milk 61 and butterfat production and butterfat percentage, when crossed with the various dairy breeds. An attempt was also made to observe if any heterosis had occurred in the first crossbred generation. Crossbreeding the Dairy Breeds The success of Mendelism, after its rediscovery in 1900, in explaining satisfactorily the mechanism of the inheritance of the more distinct physical differences between animals led to the setting up of several crossbreeding experiments in dairy cattle. Wriedt (1930) reported what is perhaps the earliest crossing between two dairy breeds-«begun in Denmark in 1906 in a herd containing pure— bred Red Danish Milkrace and purebred Jersey cattle. Jersey bulls were used on both Red Danish Milkrace and Jersey cows. The ob— ject of the experiment was partly to analyze the genetic control of production characteristics, particularly that of butterfat percentage. The investigations comprised a total of 1,175 cows. Included were the Red Danish Milkrace dams, their purebred daughters, the purebred daughters from the Jersey bulls used in crossing back, and all the crossbred cows of various degrees of crossbreeding. Table I shows the results of this experiment reported for 70—day 62 TABLE I (From Wriedt, 1930). u..- _ - .___.__ Milk Yield Fat Yield F 1: Breed in First (:0;th in First 70 Days (0,) 70 Days (lbs.) '3 (lbs.) Red Danish Milkrace (175) 1,694 3.40 67 F1 (71) 1,574 4.39 80 Purebred Jersey (116) 1,350 5.57 87 ”— — -.——-———._ __ __.— -——__- .. _. _. .-_..- ——...- yields of milk and fat content in the first lactation. The number of animals involved in each group makes the data very meaningful. The Jersey purebreds used for comparison to the Jersey parental breed were daughters of the bulls that sired the F1 and the Red Danish Milkrace are the dams of the F1. The means of the parental breeds are 1,522 pounds of milk, 4.48 percent butterfat content, and 77 pounds butterfat yield, respectively, for the seventy days observed. The F1 yield is very close to the means of the parental breeds for all these characters. Reciprocal crosses showed no consistent differences . 63 The author concluded that one genetic factor was respon— sible for the difference between the two breeds, basing such con- clusion on the butterfat percentage in the milk of Jersey and Red Danish cattle and 108 F1 hybrids, 42 backcrosses to Red Dane bulls, and 49 backcrosses to Jersey bulls. The gene for high fat content had an equal effect whether it was united with a gene for high or for low fat content. There were also indications that some Jersey bulls carried a modifying gene for increased fat percentage. In 1911, crossings between two dairy breeds were begun by T. J. Bowlker in Massachusetts——a private breeder mating Guern— seys and Holsteins reciprocally. Castle (1919) reported that the object of this experiment was to obtain females with capacity to produce as much milk as the Holsteins and have the high test of the Guernseys. In 1919 the entire project was turned over to the University of Illinois. At this time, twenty—three F1 Guernsey X Holstein and eight of the reciprocal cross had finished records. The fat production for these two groups was 265 pounds and 287 pounds. respectively, as compared to 261 pounds for twenty—four purebred Holstein dams and an estimated 231 pounds for eight pure—- bred Guernsey dams (the national average for the breed at this time). While the pmduction of fat for the crossbreds is greater 64 for each group than for either parent, the significance of the dif— ferences has not been determined. Gilmore (1952) analyzed these available data in terms of Fat Corrected Milk and found the calcu— lated amounts to be intermediate between females. Although addi— tional reports were published concerning this experiment——Yapp (1929, 1930, 1931), the data contained therein were inconclusive and unsuitable for comparisons between crosses and parental breeds. Schmidt (1948) made a study of the crossing of Black Pied Lowland cows (Friesian) and Jersey bulls. The F1 generation sug— gested strongly that hybrid vigor (heterosis) actually was found in this cross. However, there was some suggestion that the Friesian cows selected for this experiment had a rather low butterfat per— centage. Table II shows the production values for the F1 cross as compared to the parental breeds. The Friesian cows were the dams of the F crossbreds and the Jerseys were a group of unrelated cows under the same management. The four Jersey bulls that sired the F crosses were selected from high butterfat percentage 1 dams. The F1 cross lies between the parental breeds in both milk yield and butterfat percentage, but sufficiently above the mean in both cases to be considerably above that of either parent breed for fat yield. Although these results approach the definition of TABLE II (From Schmidt, 1948) ._. ____._ _.__._._..._.. - —- —-- _ .— —— _ ._...—____..__ ~._—-— .__- _-_- ._.— *— ——.. _.__——- ——-— 3O 5—Day Yield Fat Fat P L t Breed First Lacta— Content 1:21:61!) 21:0,?58 Yield tion (lbs.) (%) ’° ’° (lbs.) Friesian (12) 6,527 3.12 3.11 4.50 237 F1 (12) 6,166 5.05 3.37 4.66 362 Jersey (19) 4,555 5.65 3.86 4.60 293 .——.—-—— heterosis, as defined by the American Society of Animal Produc— tion Committee on Investigations (1942), the number of animals was so small that it could well be due to that of random chance. The Bureau of Dairy Industry at Beltsville, Maryland, ini- tiated the most extensive experimenis to date in crossing dairy breeds. Reed (1945) reported that four dairy breeds were included in the experiment—~Holsteins, Jerseys, Guernseys, and Red Danes. All the foundation animals, both males and females, were of known producing and transmitting ability, and it was felt that blending this proved stock would bring forth any hybrid vigor that might result from crossing breeds . 66 The major plan of the experiment differs from the usual pattern of crossbreeding, in that it calls for continuous introduction of new genes through the use of proved sires of the respective breeds. The females resulting from the mating of two breeds —-Holstein and Jersey for example——were mated to the Red Dane sire for the three—breed crosses. The resulting three—breed fe— males in this case were then mated to either a Jersey or Holstein proved sire in the second round of the three breeds involved. Pre— liminary reports by Reed (1946, 1947, 1948, 1949, 1950), La Master e_t_ El. (1950), and Fohrman (1941-3, 1947) show that the two~breed cows produced more than their straight—bred dams. Results reported by Reed (1951) show that the three— and four—breed cows have produced slightly more milk and butterfat than their two—breed dams. All crossbred cows were tested at Beltsville as first-calf heifers, and were milked three times a day for 365 days. After completion of their first lactation rec— ords, they were sent to cooperators' herds for further study under more practical farm conditions, where they were milked twice a day for 305 days or less for several lactations. Under first—calving conditions-«at approximate average calving age, the two—breed cows produced on the average 2,868 pounds of milk and 143 pounds of butterfat more than did their 67 purebred dams. The 81 three- and four—breed cows tested to date at Beltsville averaged 290 pounds of milk and 7 pounds of butterfat more than did their two—breed dams. Under farm condi— tions, however, the three—breed cows continued to produce more than the two-breed cows. Table III shows an even more striking differ— ence between the two crosses. Even though these data for any one lactation period were produced at different years of environ- mental conditions, it is no doubt safe to assume that the second lactation of the threenbreed cows was made the same year as the fourth lactation of the two—breed cows-~their dams. Not taking into consideration the advantage of two years and three months of age that the two—breed cows have over the three—breed cows in comparing the fourth and second lactations of the two crosses, respectively, the three—breed cows have produced 995 pounds of milk and 33 pounds of butterfat more than the two-breed cows, under practical farm conditions. Robertson (1949), appraising the crossbreeding experiments with dairy cattle, concluded that while all the conditions of an ideal experiment had not been met by the Beltsville station, some logical conclusions could be drawn. Gilmore (1952) concluded that if it could be assumed that the dams of the thirty—eight crossbred cows 68 TABLE III AVERAGE PRODUCTION OF THE BELTSVILLE CROSSBRED COWS IN FIVE COOPERATOR HERDS (actual production 2 X 365 days or less) (From Reed, 1951) —__———.—~ —__-_._.__— ——.—_——————-_._.—- . ._————__— _ .. __ C ws Butterfat 0 Lactation Milk ' A Tested Percent Pounds Calvmg ge Two—Breed C0333 Second 39 10,319 4.39 449 3 yrs., 7 mos. Third 33 10,781 4.38 466 4 yrs., 6 mos. Fourth 26 10,544 4.42 460 5 yrs., 9 mos. Three—Breed Cows Second 35 11,539 4.31 493 3 yrs., 6 mos. Third 21 11,758 4.3-6 511 4 yrs., 8 mos. are relatively similar in genotype to the dams of the forty—one purebred cows, the difference of 120 pounds of butterfat in favor of the crossbreeds is strikingly significant as reported by Fohr— man (1949). This seems to be a heterosis effect of about 15 to 25 percent. Similar results with fat—corrected milk were obtained 69 by La Master e_t_ El; (1950) with fewer cows. If this temporary conclusion on the appearance of heterosis is substantiated, it will be the first clear—cut case, with large numbers recorded, in which dairy cattle are involved. Few experiments have been planned to show ”grading up." However, this type of breeding has been practiced by thousands of breeders who use a succession of registered bulls. After the suc- cessive use of this system, grade cattle become indistinguishable from purebred cattle in all measurable genetic characteristics. Still they are ineligible for registration in closed herdbooks. What was perhaps the first plamied grading—up experiment was started in 1907 at Iowa State College. The object of this ex— periment was to determine the influence of a succession of pure— bred sires on the production of progeny from foundation scrub cows, as well as the effect of improved feeding and management. Mc~ Candlish gt a]... (1919) reported a preliminary analysis of this ex— periment. Improvement of the environment of the foundation cows —-they were moved from Arkansas farm conditions to the conditions of the college herd—“caused an increase in production of 14 percent 70 more milk and 8 percent more fat for mature cows. Those coming in the herd as heifers increased 27 percent in milk and 24 percent in fat production. Three cows——sired by scrub bulls which were either purchased as calves or dropped on the farm——showed an increase of 10 percent in milk and 13 percent in fat. The authors concluded this was due to the more favorable surroundings. First—generation grades sired by purebred bulls of Holstein, Guernsey, and Jersey breeds showed an average increased production of 39 percent in milk and 35 percent in fat yield. Still the first generation Guernsey grades were quite variable; one decreasing 31 percent in milk and 23 percent in fat yield under her scrub dam. The Holstein and Jersey crosses increased production in all in— stances. The second—generation grades——when averaged irrespec— tive of breed of sire——showed an increase of 130 percent in milk and 109 percent in fat over the scrub foundation cows. The persis— tency of production of the grades increased progressively with the gene ration number. The investigators concluded that grading up of scrubs through the use of purebred sires increased production. Weaver _e_t _a_l__. (1928) reported a further analysis of this ex— periment. Although the number of cows completing records during the interim of the ten years was relatively small, findings for the 71 first and second generation grades were essentially the same as those in the earlier report. These results of the first and second generation grades established the value of purebred sires in grading up a herd from common stock. However, it was deemed desirable to continue the grading up with at least one of the breeds, to de— termine how far the grading up might be continued before a point was reached beyond which increases would not be easily secured. The matings were carried beyond the second generation only The production of four third-genera- in the case of the Holsteins. tion Holsteins averaged 139 percent greater than that of the scrub cows. They manifested an increase of only 6 percent over their second—cross dams. There were only three grades of the fourth generation to complete records of which the authors take cognizance All three produced less than their third— as being small in number. gene ration dams——-an average decrease of 15 percent. The authors therefore concluded that a point in production had been reached with the third—generation Holstein grades beyond which further improve— ment could not be regularly anticipated. Olson and Biggars (1922) reported the results of grading up of grade beef animals by use of purebred dairy sires at the South Dakota station. Their results, consisting of first and second 72 generation grades, are in close harmony with those of the Iowa experiment. Skvorcov and Basov (1940) reported the results of grading up of Siberian grade dairy animals with purebred Simmental bulls. The performance of 93 first— and second-generation crosses yielded an average increase in production of 38.1 percent in milk and 25 percent in butterfat over the foundation dams. Garkavi (1945) ap— praised a grading-up plan aimed at accelerating the production ef— ficiency of cattle on collective and state farms in Russia. He es— timated it would take twenty—five to twenty—nine years for a herd of local unimproved cattle to reach a graded—up point of fifth—gen— eration animals. He found, according to the 1939 census, that 84.5 percent of the cows in the collective and state farm herds were first—generation--—or more——crosses. In fourth—generation East Friesian and Siberian crosses the butterfat percentage had fallen from 4.2 percent to 3.2 percent, while in fourth generation Sim—— mental and Siberian crosses it was 3.9 percent. A report on the attempts of grading up native Brazilian cattle with European dairy breeds has been made by Rhoad (1943). The first—generation crosses-«resulting from using European dairy bulls on native cattle like the Caracii and Crialo——increased in 73 milk production. However, further grading up tended to result in inferior animals progressively less well adapted to the environmental surroundings of that region. Valerani (1951) found that the use of Friesian bulls on grade Brown Swiss cows in Italy increased pro- duction. Over 15 years, the respective milk yields of the crossbred ., F F and F cows were 27.6, 35.3, 41.8, and or graded-up F1 2, 3, 4 50.0 percent more, respectively, than the original Brown Swiss dams. Repeatability of Milk, Butterfat, and Butterfat Percentage Introduction Repeatability may be defined as the regression of future performances of phenotype on past performances as measured by the expression of one trait. Many of an animal's important characteristics vary in their phenotypic expression from one time to another——thereby reducing the repeatability value of that trait. Most of the variations from one time to another are due to variations in the environment which prevail during the expression of the character. A knowledge of repeatability of economic characteristics in modern animal 74 breeding and selection techniques has become a tool of considerable value, since it is impossible to correct for incomplete repeatability due to environmental factors. Many correction factors have been devised to correct characteristics so that they have more complete repeatability, but the law of diminishing returns usually makes it scarcely worth while to use more than two or three of the most important correction factors. The purpose of this study was to determine the repeatability estimates for milk and butterfat production of the American Red Danish cattle population. These estimates were necessary in pre— dicting the real producing ability of a cow. It is important that the Red Danish breeders know these statistics for use in their breeding operations . Methods Fo r EstimatinLRepeatability Partial correlation. Stewart (1945) used the partial cor—- relation method for estimating repeatability of prolificacy in swine. Since repeatability (R) is defined as the regression of future per— formance of phenotype as measured in one expression of a trait, it may logically be estimated by the regression of the second record 75 on the first. Prasad (1951) explained the equations used in re- gressing one record on another as follows: Let (1) and (2) denote the first and second records, respec— tively, then the regression coefficient is 2 Z Z = . Z . ' b21 [Cov 121/ 1 where Cov 12 estimates O'y + cc , and Z 2 2. Z 2 El estimates crp such that b21 = [O'Y + 0c ]/0p and b21 = R. The correlation coefficient of the first and second record f2”? , is. r21 - [Cov.21]/ >32 21 . Prov1ded there has been no . . Z 2. . . selection, the expectation of 22 and 231 is identical, and thus 2 2 Z r21 = [cry + co ]/crp , which is equivalent to the regression procedure. Analysis of variance. Dickerson (1940), Briquet and Lush (1947), and Chandrashaker (1951) have reported the use of the analysis of variance method for determining repeatability. The formula, R = 2 . . . . 2 . . . . 2 [22 Within individuals] + [2 within 1nd1V1duaIS + 2 between in- 2 , , 2 dividuals], where 2 within indiViduals and 2} between individuals are the variances or mean square components of an analysis of variance, was used by the above investigators to determine an es— timate of repeatability between successive phenotypic expressions of a trait. 75 on the first. Prasad (1951) explained the equations used in re— gressing one record on another as follows: Let (1) and (2) denote the first and second records, respec— tively, then the regression coefficient is Z Z Z b21 == [Cov.12]/Z?1 where Cov. estimates O'y + crc , and 12 Z 2 2 Z 2 El estimates up such that b21 = [cry + 0c ]/up and b21 = R. The correlation coefficient of the first and second record 1 - — [c ]/,/2 27582 P ’d (1 th h b s. r2,1 - ov.21 2 1 . rov1 e ere as een no . . Z 2 . . . selection, the expectation of 22 and 21 is identical, and thus 2 2 2 , , , r21 = [try + (rc ]/0"p , which is equivalent to the regressmn procedure. Analysis of variance. Dickerson (1940), Briquet and Lush (1947), and Chandrashaker (1951) have reported the use of the analysis Of variance method for determining repeatability. The formula, R = [22 within individuals] + [>3Z within individuals + 22 between in— dividuals], where 22 within individuals and 22 between individuals are the variances or mean square components of an analysis of Variance, was used by the above investigators to determine an es— timate of repeatability between successive phenotypic expressions ( of a twait. 76 Intraclass correlation. This method—~permitting the use of all the records by each individual (Snedecor, 1950)——was demon- strated by Lush and Molln (1942) and Prasad (1951). is found such 2 Z 2 that r1 = [m1 — m2] 4 [m1 + (K - 1)mz] = [tri ] + [a + 61 ], The intraclass correlation coefficient, r1, where m1 is the mean square between individuals, m2 is the mean square between records by the same individual, and K is the number of records per individual. However, if the number of records for all individuals is not the same, K is something less than the mean number and for the intraclass correlation is found by the formula: KO — [l/(n — 1)][ZK — (EKZ/EKH. Estimates of Repeatability Linfield (1900), in a study of production records at the Utah Station, published what is probably the first attempt to esti— mate repeatability of production characteristics of dairy cattle in the United States. He concluded that the average of several years' production was the most reliable index of a cow's productive ca— pacity, since there is a large variation between the records of the same coW. Gavin (1913b) contributed to the early work in estimates 1 77 of repeatability and found. when studying over two thousand Holstein— Friesian and Shorthorn production records, that the correlation co— ‘ efficient between successive records did not rise above 0.6. Since those early works were published, numerous reports have been published giving estimates of repeatability of production records in dairy cattle. For the sake of brevity, these have been set out in tabular form in Table IV. ' Heritability of Production of Milk, Butterfat, and Butterfat Percentage Intro duction Heritability can be defined as that fraction of the observed or phenotypic variance which is caused by differences between genes or the genotypes of the individuals in a particular population (1411311, 1949). Under this definition it can be used in both a broad 1 a-_.-1 a narrow sense. In a narrow sense, the heritability of a char— aCteristic refers to the degree to which observed variation in that ) cha~1‘acteristic is due to the average effects of the genes carried 1 by eaeh individual. In a broad sense, the heritability of. a charac— tel'13tic refers to the functioning of the whole genotype as a unit in contrast to the environment. However, according to the Mendelian -' -U~—-v—.-—u...4—- .- 1 78 TABLE IV ESTIMATES OF REPEATABILITY FOR PRODUCTION CHARACTERISTICS IN DAIRY CATTLE Repeat- Method Used to Determine ability Repeatability Rderence I. Repeatability: A. Milk 0.60 Correlation between records Gavin (1913b) (Holstein-Friesian, Shorthorn 7 day) 0.67 Intrabreed correlation Gowen and Gowen (1922) (Holstein-Friesian A. R.) 0.73 Within-herd correlation Sanders (1930) (Corrected for all environ— ment factors) 0.52 Within-herd correlation Sanders (1930) (Uncorrected records) 0.50 Intraherd correlation Dickerson (1941) (Correction of Gowen, 1922) 0.26 Intraherd correlation Dickerson (1940) (Uncorrected for age and calving interval) 0.35 Intraherd correlation Dickerson (1940) (Corrected for age and calving interval) 0°55 Intraherd correlation Gowen (1924b) (Guernsey) 0'50 Intraherd correlation Gowen (1934) (Jersey) TABLE IV (Continued) Repeat- ability Method Used to Determine Repeatability Reference 0.50 0.48 0.29 0.34 0.40 0.44 0.45 0.49 0.12, Intrahe rd correlation (Holstein—Friesian A. R.) First record over the later records Intraherd correlation (Holstein-Friesian 305 uncorrected) Intraherd correlation (Holstein-Friesian 305 herd test, 2X.) Intraherd correlation (Holstein-Friesian 305 2X. M. E. Association Factors) Intraherd correlation (Holstein—Friesian 305 2X. M. E. calculated factors) Analysis of Variance (Holstein-Friesian) Analysis of Variance (Jersey) Analysis of Variance (Guernsey) Analysis of Varianc e (Ayr shi re) day, day, day, day, Verna (1945) Lush and Strauss (1942) Laben and Herman (1950) Laben and Herman (1950) Laben and Herman (1950) Laben and Herman (1950) Chandrashaker Chandra shaker Chandra shake r Chandra shaker (1951) (1951) (1951) (1951) TABLE IV (Continued) 80 Repeat- Method Used to Determine R f ability Repeatability e erence 0.70 Analysis of Variance Chandrashaker (1951) (BrOWn Swiss) 0.49 Analysis of Variance Chandrashaker (1951) (Combined 5 breeds) II. Repeatability: B. Butterfat .0.6O Intrabreed correlation Gavin (1913b) 0.69 Intrabreed correlation Gowen (1920c) (Jersey A. R.) 0.71 Intrabreed correlation Gowen (1924a) (Holstein—Friesian A. R.) 0-55 Intraherd analysis of Harris _e_t a_l. (1934) variance (C. T. A. Records) 0.33 Intraherd analysis of Harris _e_t_ a_l. (1934) variance (Random records from same herd) 0.33 Intraherd analysis of var- Harris e_t 31. (1934) iance (Average between records of same cow) 0'60 Intrabreed correlation Plum (1935) (All herds) 0.40 . Intraherd analySis of Pluzn (1935) variance 81 TABLE IV (Continued) m Repeat- ability Method Used to Determine Repeatability Reference 0.51 0.68 0.43 0.43 0.34 0.36 0.43 \ (Iowa DHIA Correlation between succes- sive records (Guernsey A. R. partial records) Correlation between succes- sive records (Guernsey A. R. 365 day records) First rec0rds over later records Intrabreed correlation (Jersey A. R.) Intrabreed correlation (Jersey H. 1. R.) Intrabreed correlation (Holstein-Friesian H. 1. R.) First records over later records Intrahe rd correlation Intrahe rd cor relation (Swedish records) First and later record correlation records) Gaines (1936) Gaines (1936) Lush and Arnold (1937) Copeland (1938) Copeland (1938) Berry and Lush (1939) Lush (1940) Dickerson (1940b) Johans son and Hanson (1940) Lush and Strauss (1942) f ._....... ___..——._..—____ ._ 82 TABLE IV (Continued) Repeat— ability Method Used to Determine Repeatability Reference 0.40 0.46 0.50 \ First and later record correlation (Holstein H. I. R. records) Intraherd correlation (8 month Holstin A. R. records) Analysis of variance (Holstein-Friesian A. R. noncontemporary records) Analysis of variance (Holstein—Friesian A. R. contemporary records) Analysis of variance (Holstein-Friesian H. 1. Analysis of variance (Holstein—Friesian H. 1. Analysis of variance (Jersey H. I. R.) Analysis of variance (GuernSey H. I. R.) Analysis of variance (Ayrshire H. I. R.) Analysis of variance (Brown Swiss H. I. R.) Analysis of variance (Five breeds combined) Lush and Strauss (1942) Verna (1945) Laben and Herman (1950) Laben and Herman (1950) Chai (1951) Chandrashaker (1951) Chandrashaker (1951) Chandra shaker (l 951) Chandrashaker (195 1) Chandrashaker (195 1) Chandrashaker (19 51) 83 TABLE IV (Continued) Repeat- ability Method Used to Determine Repeatability Reference 0.52 0.72 0.66 0.76 0.61 III. Repeatability: C. Butterfat Percentage Intrabreed correlation (Jersey A. R. records) Intrabreed correlation (Holstein-Friesian A. R. 365 day records) Intraherd correlation (Holstein-Friesian 8 month A. R. records) Analysis of variance (Noncontemporary Holstein A. R. records) Analysis of variance (Contemporary Holstein A. R. records) Gowen (1920a, 192 Ob) Gowen and Gowen (1922.) Verna (1945) Laben and Herman (1950) Laben and Herman (1950) 84 laws of segregation and recombination of genes, it is impossible that the genotype as a unit be transmitted. Instead, some of the genes may interact with others in such a way as to produce a nonadditive effect, which in certain combinations has effects quite different from their average effects in a given population. 'i'he differences between actual effects in each combination and their av— erage effects in the whole population are called dominance deviations and epistatic deviations, which are seldom, if at all, transmitted. Since they would not materially add to one's estimate of heritability of characteristics, these nonadditive influences are generally dis—— counted. A well—recognized and almost world—wide desire by breed—- 31'3 0f dairy cattle for information of any kind concerning the hereditary basis of the production characteristics has led to the Conduct of an exceedingly large number of varied investigations. The PrOblem has been tackled from many angles, and results have been PUblished in almost every conceivable manner. Before deal— ing With the data of more recent date, mention should be made of earlier views on the inheritance of production characteristics. The breeding practices of former times are of extreme interest—~es- e P “la-11y since many of them still prevail. It cannot be denied that those old—time practices won for Britain her reputation as the stock farm of the world (Smith and Robison, 1933). At that time, breeding practices were not attuned to the principles of heredity, for the science of genetics was not yet born. The principles of the old—time breeders, such as Bakewell, Cruickshank, and Bates, may be summarized as follows: 1. Conception of the desired type. 2. Mating the best to the best. 3. Appreciation of the pedigree value accompanied by the intelligent application of inbreeding. 4. Use of the progeny test. The views of the Pre—Mendelian Period are described in detail by Patow (1926), and therefore will be enumerated only briefly_ In this period, there were three determinants of the meth— Ode of breeding dairy cattle. First, there was formalism; secondly, a conception of heredity based upon Galton's law of ancestral in— heritance; and finally, milking tests which were concerned only with absolute yield. Galton was the first to employ ancestral correlations as a means for the study of heredity resemblance. Rietz (1909) was th e fix-St to apply it to the problem of the inheritance of a production 86 characteristic in dairy cattle, namely, butterfat production. He found the correlation between daughter and dam records to be of the order of 0.30 for Holstein—Friesian cattle. The data used were seven—day "official tests" published in Volumes 11 to 18 of the Holstein—Friesian Advanced Register. It was left to Gowen, how— ever, to make a comprehensive study of the problem of inheritance of butterfat production. Others—~especially those of the University of Missouri under Turner, and those from Iowa State under Lush—- have subsequently extended the work to all production phases and modified the methods as the requirements of the material neces— sitated. It is necessary that breeders know the estimates of herita- bflitY for characters for which they are selecting. These statistics {01‘ the American Red Danish cattle population are unknown. The PurPOSe of this study was to determine those statistics. Wor Estimating He ritability Depending on the method used, an actual numerical estimate of heritability is usually between the narrow and broad definitions. al most always including a little of the epistatic variance and some— tin: e8 a. little of the dominance variance. It may include all, part, 87 or none of the variance caused by the nonlinear or joint effects of heredity and environment (Lush, 1948). Isggenic line method. An isogenic line is a group of indi—- viduals having the same genetic composition. The only example of isogenic lines in dairy cattle is identical twins. However, since the frequency of identical twins in cattle is low, the method is of not much practical importance for the study of heritability of traits in dairy cattle. Long inbred lines do, however, approach the con- dition of the same genetic composition (Lush, 1948). The variance between members of the same isogenic line is wholly environmental. Since the relationship between members of an isogenic line is 100 percent, the ratio of the variance or the correlation coefficient is the estimate of heritability. Therefore, the heritability estimate could be obtained in two ways: (1) the method of intraclass correlation, where the variance between the isogenic: lines and that within the lines is compared—~giving a direct estimate of heritability, and (Z) the method of single clas— Sifica‘tion of analysis of variance presents another approach. If we can that variance within isogenic lines W and use B to repre— sent the variance between unrelated individuals in the population und er study. heritability can be computed simply as [B — WVB- 88 This method of estimating the heritability of traits in dairy cattle is becoming of Inore importance, since numerous sets of twins are being collected by many experiment stations——particularly those of New Zealand and Sweden. Hancock (1949) of New Zealand has estimated that one set of identical twins replaces twenty pairs of randomly selected animals. Since the New Zealand workers have collected over two hundred pairs of identical twins they can replace over four thousand animals selected at random without loss of stati stical efficiency. Regression of offspring on mid—parent. Nelson (1941) dem— onstrated the mid—parent offspring correlation or regression method to estimate heritability. Since an average of both parents is needed in the place of one, the literature on this method—~as in most other methods—45 limited. Most of the studies of heritability——and es— Pecially those in dairy cattle——have been of such nature that only one parent exhibited the character that could be directly measured. HOWeVer, this method may be used where the trait can be mea— sured in both parents-~such as fleece—weight in sheep, weight at six months of age in hogs, and pounds gain per unit feed by beef cattle 89 Correlation of parent and offspring. This method of cor— relation was first advanced by Galton and used by Rietz in 1909. Following the work of Rietz, many workers used this method to compile voluminous data on correlations between dam and off— spring. The parent—offspring method is not applicable when one of the parents does not exhibit the characteristics desired to be stud- ied. Resemblance to the sires is not very helpful in data in which the sires are fewer than the dams and, therefore, have been se— lected intensely on their own phenotypes. In random mating popula— tions, the phenotypic correlation between a parent and an offspring, , Z 2 Z 2 2 as described by Lush (1948), 15 [1/2] [(60 /0’P ) + e ree + i h r11]. 2 If the e ree term can be measured and removed, simply doubling the remainder will give an estimate of heritability. i If mating is consanguine, the ICC term becomes : [(1 + tin/Hm, instead of simply 1/2. If dominance and ePistasis were zero, heritability could be estimated by first 1"an-“cfi’ing the error term, doubling, and dividing by [1 + m]. If mating is assortive by phenotypes the same procedure 1 s followed, except that division is by [1 + r ]. sd 90 Intrasire dam—daughter regression or correlation, Esti— _mates most commonly available during the last decade are those based on this method. Lush (1940) discussed the methods of intra— sire regression and correlation. In the data where the dams are much more numerous than the sires——as in dairy cattle, this method offers a very useful way of estimating heritability. This device dodges corrections for the mating system, because the differences being investigated are only those which exist between females mated to the same sire. It also dodges many of the difficulties about correcting for the environ— mental term in the regression or correlation. The regression is interpreted in the following manner. Let X and Y represent records of offspring and dam, respectively. Then the regression coefficient, bxy' is found as follows: bxy = [Cov. XY] + ZzY. Where Cov. and ZZY are the estimated covariance and vari— ance, respectively, as described by Snedecor (1950). Gov. XY es— timates [1/2]0'2g and 22 estimates crzp so that bxy = [(1/2)0’Zg] + “2P, and H = 2b xy° The correlation coefficient is found as follows: r = __ __ xy [C°"- XY] -:— szx 223'. 91 For the case where V——the average of n records for each dam—-is used in the computation of the intraregression of offspring on dam, H is derived at from this regression in the following man— ner: H = be [(1 + [n — 1]R) + n] (Lush and Strauss, 1942). Av- eraging records for the offspring has no effect on the deviation of H. Half—sib correlation. Baker, Hazel, and Reinmuller (1943) have demonstrated that the sire component of variance is a handy tool with which to obtain the intraclass correlation for estimating heritability. However, unless very large amounts of data are avail— able (which would substantially reduce the errors of sampling) this method is not as accurate as other methods previously described. Data that are analyzed by the intraclass correlation analysis °f Variance, as outlined by Snedecor (1950), give the required sta— tistic. This, multiplied by four, gives an estimate of heritability. It is necessary to multiply by the factor, four, since the coefficient of genEtic relationship between half—sibs, if randomly bred, is 0.25. Full—sib correlation. This method——more accurate than the h ‘ . alt 3113 method——is used less in dairy cattle because of the relatively few f“lllwsibs to be found in a population. In swine and sheep this me a‘BLi-Il‘e could be more readily used. 92 It is very similar to the parent-offspring correlation method. The coefficient of genetic relationship between full-sibs in a. randomly bred population is 0.50; therefore, the regression or correlation co— efficient is multiplied by two. liistimates of Heritability In view of the variety of the material and the methods by Which they have been secured, no attempt has been made here to review the full significance of the figures. These results are tab— ulated in Table V in as convenient a manner as possible, in an attempt to summarize the investigations. There are numerous other data pertaining to the estimate or heritability of production characteristics. Smith and Robison (1931) Present an extensive list of correlations between milk, fat, and fat percentage of various pairs of relatives computed by many different investigators. The phenotypic correlations between rela- tives more remote than half—sibs must be multiplied by factors so large that it magnifies the experimental error greatly, and, in the second place, environmental errors are nearly impossible to Correct for, since undoubtedly the expressions correlated were ex— r P ease-d at widely separated intervals. Therefore, these data are TABLE V ESTIMATES OF HERITABILITY OF PRODUCTION MILK, FAT, AND FAT PERCENTAGE Chara¢_ te ristic Milk Milk Milk Milk He rit— ability E sti— mate Breed Inve stigator Yield Yield Yield Yield Yield Yield Yield Yield Yield Butte rfat Yield B utte If at Yield Butte rfat A. Regression of Daughter on Dam 0.45 0.42 Holstein— Friesian Guernsey Red Poll Ayrshire Shorthorn Jersey Holstein— Friesian All breeds Ayrshire Holstein—- F rie s ian Holstein— Friesian Red Poll Gowen (1924a) Gowen (1926) Agar (1926) Smith e_t 311.- (1931) Edwards (1932) Gowen (1934) Lush and Strauss (1942) Lush and Strauss (1942) Gifford and Turner (1928) Rietz (1909) Hills and Bolland (1913) Agar (1926) \ 1 i ‘ ‘1 (. 1' I t , ‘ ‘1 1 11 S i A ~ 1 TABLE V (Continued) \ Herit- Cha. rac- ability B reed Inve stigator teristic Esti— mate Butte rfat 0.26 Ayrshire Gifford and Turner (1928) Yield Butterfat 0.23 Guernsey Gifford (1930a) Yield . Butte rfat 0.40 Jersey Copeland (1931) Yield Butterfat 0.40 Jersey Gowen (1934) Yield Butterfat 0.12 Holstein— Plum (1933) Yield Friesian Butterfat 0.28 Holstein— Lush and Arnold (1937) Yield Friesian Butterfat 0.25 Holstein— Lush _e_t a_1_ (1941) Yield Friesian Butterfat 0.39 Holstein— Gowen (1924a) ercentage Friesian Butterfat 0.30 Guernsey Turner (1925) e1‘centage Butterfat 0.42 Guernsey Gowen (1926) e1‘Qentage Butterfat 0.35 Jersey Turner (1927b) ercentage 95 TABLE V (Continued) Herit— Cha rac — ability te ristic Esti— Breed Investigator mate Butterfat 0.33 Red P011 Agar (1926) Pe rcentage Milk Ivi ilk Milk _B. Intrasire Daughter—Dam Correlation and Regression Yield Yield Yield Yield Yield Yield Yield B uthe rfat Yield Butte rfat Yield Butt"~‘1-fa1: Yield Butte rfat 0.33 Holstein— Friesian 0.36 Holstein— Friesian 0.06 Ayrshire -" .27 Holstein- Friesian 0.26 Jersey —0. 10 Guernsey —0.22 Brown Swiss 0.25 Holstein— Friesian 0.28 Holstein- Friesian 0.30 Swedish 0.33 Holstein— Friesian Lush §_t §_l_. (1941) Laben and Herman (1950) Chandrashake r (1951) Chandrashake r ( 1951) Chandrashaker (19 51) Chandrashaker (1951) Chandrashaker ( 1951) Lush and Shultz (1936) Lush (1940) Johansson and Hansson (1940) Lush _e__t _a_l_. (1941) _——— .— sum—— 96 TABLE V (Continued) He rit— Cha rac - ability B In te ristic Esti- reed vestigator mate Butterfat 0.17 Holstein— Lush and Strauss (1942) Yield Friesian (single records) Butte rfat 0 .27 Holstein— Lush and Strauss (1942) Yield Friesian Butte rfat 0 .29 Holstein- Laben and Herman (1950) Yield Friesian Butte rfat 0.31 Holstin— Chai (1951) Yield Friesian iytte rfat 0.17 Holstein— Chai (I. 951) 131d Friesian B (single records) Yiutte rfat 0.17 Holstein— Chandrashaker (1951) eld Friesian Bf‘tte rfat 0.37 Jersey Chandrashaker (1951) Yield B “Niel-fat 0,05 Guernsey Chandrashaker (1951) Yield B t u te rfat 0.21 Ayrshire Chandrashaker (1951) Yield But te rfat —0.26 Brown Chandrashaker (1951) Yield , . Sw1ss \ 97 TAB LE V (Continued) Herit— Cha rac —- ability , te ristic Esti— Breed Investigator _ mate C. Full—sib Correlation Milk 0.55 Holstein— Gowen (1924a) Production Friesian Milk 0,41 Guernsey Gowen (1926) Production Milk 0,39 Jersey Gowen (1927) Production Milk 0.46 Ayrshire Smith and Robison (1931) Produc tion £411}: 027 Holstein— Laben and Herman (1950) roduc tion Friesian Biltte rfat 0.478 Ayrshire Smith and Robison (1931) Yield gutterfat 0.39 Holstein— Laben and Herman (1950) ”Id Friesian gutterfat 0.46 Holstein— Gowen (1924a) ercentage Friesian B Putte rfat 0.44 Guernsey Gowen (1926) ercetItage Bu P ttel‘fat 0.41 Jersey Gowen (1927) ercentage Bu 9:“ rfat 0.46 Holstein— Laben and Herman (1950) ceIntage Friesian \ TABLE V (Continued) 98 Herit- Cha rac -— ability B t teristic Esti— reed Inves igator mate Half~sib Correlation Milk 0.36 Holstein— Gowen (1924a) Production Friesian Milk 0.38 Holstein— Gowen (1924a) Production Friesian Milk 0.13 Guernsey Gowen (1926) Production Milk 0.15 Guernsey Gowen (1926) Production Milk 0,23 Jersey Gowen (1927) Production Milk 0.20 Jersey Gowen (1927) Production Milk 0,39 Ayrshire Smith and Robison (1931) prod~L1ction 1:111: 03,2 Ayrshire Smith and Robison (1931) 1'oclu-ction 1:111; 0.28 Ayrshire Smith and Robison (1931) rodllction M Pnk 0,13 Holstein— Laben and Herman (1950) roduction Friesian M 111‘ 0.14 Holstein- Laben and Herman (1950) roduction Friesian \ k rix11‘ . TAB LE V (Continued) Herit— Charac— ability Breed Investigator teristic Esti— mate 1 Butterfat 0.37 Ayrshire Smith and Robison (1931) Yield Butterfat 0.56 Jersey Copeland (1931) Yield Butterfat 0.42 Ayrshire Smith and Robison (1931) Yield Butterfat 0.46 Ayrshire Smith and Robison (1931) ‘ Yield Butterfat 0.30 Holstein- Laben and Herman (1950) Yield Friesian Butterfat 0.14 Holstein— Laben and Herman (1950) Yield Friesian Butterfat 0.37 Holstein- Gowen (1924a) Percentage Friesian Butterfat 0.17 Holstein— Gowen (1924a) ‘ Percentage Friesian 1 Butt ‘ erfat 0.17 Guernsey Gowen (1926) ( Percentage Butterfat 0.19 Guernse Gowen (1926) Y Percentage B utter-fat 0,25 Jersey Gowen (1927) ercentage B . utterfat 0.20 Jersey Gowen ”(1927) ercentage \ TAB LE V (Continued) Herit— Charac— ability Breed Investigator te r1 stic Esti— mate Butterfat 0.37 Ayrshire Smith and Robison (1931) Yield Butterfat 0.56 Jersey Copeland (1931) Yield Butte rfat 0,42 Ayrshire Smith and Robison (1931) Yield Butte rfat 0.46 Ayrshire Smith and Robison (1931) Yield Butte rfat 0,30 Holstein— Laben and Herman (1950) Yield Friesian Butterfat 0,14 Holstein— Laben and Herman (1950) Yield Friesian Butterfat 0.37 Holstein— Gowen (1924a) Percentage Friesian Butterfat 0.17 Holstein~ Gowen (1924a) ercentage Friesian Butterfat 0.17 Guernse Gowen (1926) Y ercentage ‘ Butterfat 0.19 Guernsey Gowen (1926) eret‘zntage B utterfat 0.25 Jersey Gowen (1927) ereentage B . utte rfat 0.20 Jersey Gowen ’(1927) ereentage \\ ‘ 100 not presented in this review, although their presence is recog- nize d. Inve 5 ti gational P roc edu re G rading— Up Eff e c t The data utilized in this analysis have been standardized alld processed on IBM Card Number 2. In calculating the mean production and weighted means, the formula, 31' = gm) -:— N, where 52 equals the mean, g equals N val—— “38 summated, x equals the values, and N equals the total number of items in the sample, was used. In order to describe more completely the variability of the mean and how descriptive it is of the population, the standard de— viation of each mean was calculated. The formula for the standard deviation is: 0' = /[E(x _ i)2 + N; where 0' represents the standard deflation, 2(x — 32 under the square root sign equals the deviation from the mean squared, and N equals total number of items in the samp1e. However, another formula more adapted to use in IBM calculations can be used. Since the main part of the work is to find the sum of squares of the deviations, it can be shown that: 101 Z(x — 532 = 2(x2) — [S(X)]Z/N. This formula is especially useful in machine calculations (Goulden, 1949). The ”t" test was employed to test for the significance of the differences between the various means. Since each group con- sisted of different numbers of individuals, the following formula (Snedecor, 1950) was employed: _ 7 t = x/ + — 2 A + z n1n2(nl n2 ) . (n1 n2) x , Where 11? equals difference between the two means to be tested, n1 and :12 equal, respectively, the number of items in the two sam— 2 pies, and 2x equals the pooled variance (total sums of squares) 0f.the two samples to be tested. Repeatability Estimates The production records involved in this study were made by two Categories of animals: (1) mates of the sires, and (Z) progeny resulting from the succession of purebred Red Dane sires bred to the fol"Ln-(”lation cows and the resulting crosses. The method employed in estimating the repeatability of these data, was single classification of analysis of variance, intraclass correlation (Snedecor, 1950). The variance in a sample may be separat 92d into two parts——one reflecting the natural variation of 102 individuals treated alike, and the other arising from genetic or environmental conditions peculiar to the subsamples. The former . 2 . . is an estimate of 0' , which may be assumed equal in the sampled 2 populations, while the latter is an estimate of am , the added Portion relation to differences among the population means. The 2 2 sum, 0‘ + (rm , is the variance of individuals picked at random from the entire universe made up of the sampled populations. Th 2 2 2 , e ratio of the two variances, crm + (1r + (rm ), is known as the intraclass correlation, r1 = R. The justification for calling this a ”correlation" is that correlation is a two-way average of relationship (Snedecor, 1950). HoweVer, it is necessary to keep clearly in mind the character of the 1‘ela-tion being considered. A characteristic of r is revealed by diVid-ing both numerator and denominator of the ratio by (n — l): 2 2 r 2 ; lexz , l/ (23xl )(ZxZ ) = axlxz/m — 1) + 1/[Zx12/(n — 1)][zx22/(n — 1)] = Covariance + (P3 = Covariance + Geometric mean of variance. Correlation is thus the quotient of two averages of variation——one, the coVai-iance of the two measurements, x1 and x2; the other, an age of the two variances. It is an abstract number measuring V'— 103 covariation——if two related characters each occur in various sizes, their correlation is a measure of the extent to which their varia— tions are concomitant. In order to calculate intraclass correlation directly from a table of analysis of variance, it is convenient to put the for—— mula in the form: r1 = [Mi — M] —'.- [Mi + (K — HM], where Mi and M denote mean squares for between individuals and Within individuals, respectively. K is the number of items per subsample; if the subsamples differ in size, the average value, Ko’ is used. Since in these data it was desired to use all records of individual cows, K varied in size, and the formula. necessary to- find an average number, KO, may be conveniently put in the for- mula: K0 2 [1/(N — 1)][ZK — (ZKZ/ZKH, where N is the number of SUbsamples, 23K is the total sample size, and 2K2 is the total sum of squares. Since the methods used for obtaining the necessary statistics for setting up an analysis of variance table can be found in detail in numerous texts of statistics (Snedecor, 1950; Goulden, 1949; CroxtOn and Cowden, 1949), the detailed procedure for calculation Will not be presented here. All calculations—-except the several .u-u—r---H.a_':‘hq. 4._....__ -._ - - 104 involved within the analysis of variance table itself-~were performed by International Business Machines in order to reduce the errors involved in hand—operated machines. All the statistics necessary for these analyses were tabulated on the usual long, folded sheet of paper. The tabulations were made by the specific steps necessary to obtain the data from IBM Cards No. 3 and No. 4 (Fig. 3). The calculations of repeatability were made by grouping the data into three categories: (1) An intrabreed analysis was made grouping all herds together in one large population of records; the number of herds involved in each breed was different, and several of the breeds often came from the same herds. (Z) An intrabreed analysis by herds was made, and a single estimate of repeatability for each breed for all herds was then obtained (3) An intI‘abreed analysis was made grouping all herds ' get‘mr in one large population of records——keeping separate each breed and its successive Red Dane crosses to find the effects on repeatability due to a. succession of Red Danish sires. 105 He ritability Estimate 5 Of the different methods described for estimating the heritability of production traits, the following three methods were used in this study: (1) intrasire daughter—dam correlation, (2) intrasire regression of daughters on dams, and (3) paternal half- sib correlations. The following equations were used to estimate the cor— relations (r), regressions (b), and their standard error (e). Daughter on dam regression coefficient: 2 b = ny + Ex Standard error of the regression coefficient squared: Eb?‘ = [2y2 — ([Zxle/ZXZH + [(n — 2)/2x2] Dam—daughter correlation coefficient: r = in + /<2x2)<>:y2) Standard error of the correlation coefficient squared: Er2=[1-r2]—Z-[n-Z] Half—sib correlation coefficient: r1 = M}? + [M + Mx] WheI' _ e M is the mean square of the error term and Mx = [mean squar e b . emeen sires — mean square of error term] + [average numb er in each sire group]. f- ; 106 Standard error of the half-sib correlation coefficient: 2 Er1 = 1 — r + I/n —- 2 As previously pointed out, heritability was estimated by doubling the dam—daughter correlation and daughter—on—dam regression coefficients, and by multiplying by four the half—sib correlation coefficient. The respective standard errors were mul— tiplied by the same factors. Inve stigational Re suits Grading—Up Effect A general survey was made of the foundation breeds of dairy Cattle bred to the Red Danish sires in this project to deter— mine their level of production characteristics as measured by the mean, Standard deviation, and coefficient of variability of these traits; this survey served as a basic step in this study. This analysis was based on at least 90 percent of the records for each breed- An analysis of variance for milk production (Table VI) 8110 WS a significant difference between breeds (P < 0.01). The H01 stein breed had the highest milk production average followed 107 TABLE VI ANALYSIS OF VARIANCE OF PRODUCTION TRAITS De— Source of grees Sum of Mean Trait Variance Of S uares S a F Free— q qu res dom Foundation Breeds Milk Total 1,357 7,616,690,065 Between 5 1,203,272,100 240,654,420 50.7** breeds Within 1,352 6,413,417,965 4,743,652 breeds Butter— Total 1,357 10,469,884 fat Between 5 927,861 185,572 23.0”” breeds Within 1,352 9,542,023 8,058 breeds \ Red Danish X Foundation Breed Progeny Mil k Total 2,129 8,917,057,061 Between 2 24,106,000 12,053,000 2.88 Crosses Within 2,127 8,892,951,061 4,180,983 \ r—_—f 108 TABLE VI (Continued) De— Source of grees Sum of Mean Trait , of F Variance Squares Squares Free— dom Butter— Total 2,129 13,498,072 fat Between 2 606 303 0.04 Crosses Within 2,127 13,497,466 6,346 *‘9' (P < 0.01). 109 by the Brown Swiss, Mixed, Jersey, Shorthorn, and Guernsey breeds, in descending order (Table VII). The Holstein and Brown Swiss breeds were not significantly different, although they differed at the l—percent level from all other breeds (”t" test). The Mixed, Jersey, and Shorthorn breeds were not significantly differ— ent from each other, but were significantly different from the Guernsey breed (P < 0.01). The analysis of variance for butterfat production (Table VI) showed a significant difference between herds (P < 0.01). The Jersey breed had the highest butterfat average, followed by the Brown Swiss, Holstein, Mixed, Guernsey, and Shorthorn breeds, in that Order. Butterfat production was not significantly different for the Jersey, Brown Swiss and Holstein breeds, while they were different from the Mixed, Shorthorn and Guernsey breeds (”t” teSt P < 0.01). The Mixed and Guernsey breeds were different from the Shorthorn breed. Table VIII shows the production characteristics of three generations of graded-up Red Danish progeny, irrespective of breed of foundation Cow. For milk and butterfat production, the "F" test (Table VI) shows a nonsignificant value. No difference existed 110 TABLE VII THE MEAN, STANDARD DEVIATION, AND COEFFICIENT OF VARIABILITY OF MILK, BUTTERFAT, AND BUTTERFAT PERCENTAGE FOR ALL BREEDS USED IN THE RED DANE PROJECT Number Pounds Milk Breed of Cows Mean Std. Dev. I‘ C.V. Holstein 335 10,057 2,586 25.7 Brown Swiss 35 9,690 2,010 20.7 Mixed Breed 169 8,355 2,175 26.0 Jersey 149 8,335 1,803 21.6 Shorthorn 302 8,199 1,991 24.2 Guernsey 368 7,470 1,703 22.8 Weighted Avg. 1,358 8,536 2,305 27.6 Red Dane Milkrace 75 9,577 1,347 14.1 TAB LE VII (Continued) Pounds Fat Percent Butterfat Content Mean Std. Dev. C.V. Mean Std. Dev. C.V. 380 99 26.0 3.77 0.23 06.1 386 70 18.0 3.98 0.21 05.3 346 87 25.6 4.14 0.17 04.1 91 23.0 4.75 0.10 01.8 76 23.4 3.92 0.17 04.3 77 22.7 4.53 0.13 03.9 90 25.3 4.14 0.18 04.1 91 23 4.07 0.10 02.4 TABLE VIII THE MEAN, STANDARD DEVIATION, AND COEFFICIENT OF VARIABILITY OF MILK, BUTTERFAT, AND BUTTERFAT PERCENTAGE GENERATIONS OF PROGENY OF RED DANISH SIRES CROSSED WITH VARIOUS DAIRY BREEDS Number Pounds Milk Percent Increase Breed of Over C M St . D . C.V. ows ean d ev Foundation First Gen. 1,285 9,045 2,085 23.0 + 6.0 Second Gen. 710 9,197 1,939 21.0 + 7.7 Third Gen. 135 9,427 2,177 23 O +10.4 weighted 9,116 2,407 22.4 + 6.8 Average TABLE VIII (Continued) Pounds Butte rfat Pct. Butterfat Content Measn C.V. 372 373 374 03.0 02.0 114 between the first—, second—, and third—generation Red Danish fe— males for milk and butterfat production. The grading—up effect by use of three generations of pure— bred Red Danish sires on the various foundation dairy breeds is illustrated in Tables IX through XIII. By use of the "t" test for unequal numbers, the increase in milk and butterfat production that occurred in the Guernsey, Jersey, and Shorthorn progeny was found to be significant (P < 0.05 and 0.01). The crosses resulting from the Holstein, Brown Swiss, and Mixed breeds exhibited no significant change from the foundation cows. In several of the graded—up lines, the means of the crosses were significantly different from each other. Table XIV shows the results of the "t" test values for all the foundation cows and their progeny. Table XV summarizes the effects on production traits from crossing purebred Red Danish sires with the various dairy breeds in the grading—up project. The "t“ test shows the differences be— tween the mean milk and butterfat production to be highly signifi— cant, The differences in the butterfat percentage were not signifi- cant. 115 TABLE IX THE MEAN, STANDARD DEVIATION, AND COEFFICIENT OF VARIABILITY OF MILK, BUTTERFAT, AND BUTTERFAT PERCENTAGE OF HOLSTEILN—FRIESIAN COWS AND THREE GENERATIONS OF THEIR GRADED-UP RED DANISH PROGENY Pounds Milk in?” Breed Number 3::rse M S . D . . . ean td ev C V Foundation Holstein 147 9,899 2,434 24.5 First Gen. 187 9,866 2,079 21.0 -0.3 Second Gen. 59 9,116 1,966 21.5 —7.9 Third Gen. 6 9,730 1,159 11.9 -—l.7 116 TAB LE IX (Continued) Percent | Pounds Butterfat Pct. Butterfat Content Increase 0 Mean Std. Dev. C.V. ver Mean Std. Dev. C.V. Foundation 373 92 24.6 3.76 0.23 06.1 383 74 19.2 +2.7 3.88 0.07 01.8 366 so 217 —1.9 4 01 o 03 oo 7 117 TABLE X THE MEAN, STANDARD DEVIATION, AND COEFFICIENT OF VARIABILITY OF MILK, BUTTERFAT, AND BUTTERFAT PERCENTAGE OF GUERNSEY COWS AND THREE GENERATIONS OF THEIR GRADED~UP RED DANISH PROGENY Pounds Milk :e’°:nt Breed Number gilrse M t . D . . . ean S d ev C V Foundation Guernsey 147 7,586 1,649 , 21.7 First Gen. 210 8,742 2,078 23.7 +15.2 Second Gen. 83 8,923 1,956 21.9 +17.6 Third Gen. 9 10,337 2,541 24.5 +36.3 .11 I TAB LE X (Continued) 118 Pounds Butterfat I::::::: Pct. Butterfat Content Mean Std. Dev. C.V. 0"", Mean Std. Dev. c.v. Foundation 341 75 22.0 4.49 0.12 02.7 375 72 19.1 +10. 4.29 0.10 02.3 377- 86 22.9 + 9.1 4.17 0.09 02.2 416 107 25.6 +22. 4.02 0.07 01.8 ‘ or 119 TABLE XI THE MEAN, STANDARD DEVIATION, AND COEFFICIENT OF VARIABILITY OF MILK, BUTTERFAT, AND BUTTERFAT PERCENTAGE OF SHORTHORN COWS AND THREE GENERATIONS OF THEIR GRADED—UP RED DANISH PROGENY Pounds Milk :ercent Breed Number S::afse M St . D . . . ean d ev C V Foundation Shorthorn 139 8,118 1,842 22.6 First Gen. 157 8,864 1,901 21.5 + 9.2 Second Gen. 48 9,328 1,924 21.7 +14.9 Third Gen. 11 9,601 2,658 27.6 +18.3 120 TAB LE XI (Continued) Pounds Butterfat Percent Pct. Butterfat Content Increase Mean Std. Dev. C.V. over Mean Std. Dev. C.V. . Foundation 317 74 23.4 3.90 0.17 04.4 353 79 21.2 +11.4 3.97 0.07 01.8 370 78 21.2 +16.7 3.92 0.03 00.7 376 90 23.8 +18.6 3.91 0.03 00.7 121 TABLE XII THE MEAN, STANDARD DEVIATION, AND COEFFICIENT OF VARIABILITY OF MILK, BUTTERFAT, AND BUTTERFAT PERCENTAGE OF JERSEY COWS AND THREE RED DANISH PROGENY GENERATIONS OF THEIR GRADED—UP Pounds Milk 1:8:C::: Breed Number Oveer . D . . . Mean Std ev C V Foundation Jersey 88 8,212 1,740 21.1 First Gen. 138 8,690 1,691 19 4 + 5 8 Second Gen. 46 9,568 1,947 20.3 +16.5 Third Gen. 10 11,219 1 811 16.1 +36.6 TABLE XII (Continued) P Pounds Butterfat ercent Pct. Butterfat Content Increase 0 Mean Std. Dev. C.V. ver Mean Std. Dev. C.V. Foundation 382 89 23.3 4.64 0.23 05.5 1 379 79 20.7 — 0.8 4.36 0.10- 02.4 ‘ 405 72 17.8 + 6.0 4.23 0.07 01.6 430 72 16.7 +12.6 3.83 0.10 02.6 123 TABLE XIII THE MEAN, STANDARD DEVIATION, AND COEFFICIENT OF VARIABILITY OF MILK, BUTTERFAT, AND BUTTERFAT PERCENTAGE OF MIXED—BREED COWS AND THREE GENERATIONS OF THEIR GRADED—UP RED DANISH PROGENY Pounds Milk :ercem Breed Number grease ver M . D . . . ean Std ev C V Foundation Mixed—Breed 41 8,068 2,008 24.8 First Gen. 47 8,468 1,649 19.4 + 5.0 Second Gen. 28 9,049 1,433 15.8 +12.0 Third Gen. 7 8,133 2,661 32.7 + 0.8 124 TABLE XIII (Continued) Percent Pounds Butterfat Pct. Butterfat Content Increase 0 Mean Std. Dev. C.V. ver Mean Std. Dev. C.V. Foundation 349 90 25.8 4.32 0.16 03.7 361 64 17.7 +3.4 4.26 0.05 01.1 361 63 17.4 +3.4 3.98 0.05 01.3 351 68 19.3 +0.6 4.31 0.09 02.2 TABLE XIV TABLE OF "t" VALUES DETERMINED FOR TESTING THE MEANS OF MILK AND BUTTERFAT FOR DIFFERENCES BETWEEN THE VARIOUS COMBINATIONS FOR GRADING—UP EFFECT Combination Milk Butterfat De— Within Production Production grees Each of G rafiexfe—UP M 1 ”M 2 Valltliie M 1’M2 viii; 2::- Guernsey vs. F1 1,156 5.61** 34.8 4.35% 354 Guernsey vs. F2 1,337 5.37** 31.0 2.79** 228 Guernsey vs. F3 2,751 11.55** 75.0 225* 154 F1 vs. F2 181 0.689 3.0 0.300 290 F1 vs. F3 1,595 5.10“: 41.0 264* 217 F2 vs. F3 1,414 199* 44.0 241* 90 Holstein vs. F1 — 33 0.133 9.8 1.08 332 Holstein vs. F2 _ 783 234* — 7.0 0.51 204 Holstein vs. F3 - 169 0.47 2.0 0.03 151 F1 VS- F2 - 750 245* -17.3 1.5 244 F1"5- F3 - 136 0 42 — 6.0 o 31 191 F2 VS- F3 _ 414 0.501 9.0 0.031 63 Sh°rth°rn vs. F1 746 3.40“ . 35.5 4.08** 294 \ TABLE XI V (Continued) Combination Milk Butterfat De— Within Production Production grees Each of “553;”? Mt-Mz v11; Mt-Mz v3.3. 22?? Shorthorn vs. F2 1,210 6.30** 53 5.21** 187 Shorthorn vs. F3 1,483 7.42“ 59 8.26** 149 F1 vs. F2 464 247* 15.3 1.22 303 F1 vs. F3 737 1.12 23 0.23 167 F2 vs. F3 273 0.341 6.0 0.221 57 ‘JerSey vs. F1 478 204* 2.9 0.051 224 Jersey vs. F2 1,356 4.071% 23 1.98* 132 Jersey vs. F3 3,007 5.11** 48 2.70** 96 F1 vs. F2 878 2.92** 26.2 198* 182 F1 vs. F3 2,529 4.20** 51 2.86** 146 F2 vs. F3 1,651 2.49 25.1 0.29 54 BrOWn Swiss vs, Fl 67 0.012 6.6 0.05 30 Mixed vs. F1 400 1.00 12.0 0.71 84 Mixed vs- F2 981 1.27 12 0.65 68 Mixed vs, F3 55 0.11 2.0 0.13 47 581 0.356 0 0 73 TABLE XIV (Continued) Combination Milk Within Production Each Graded—Up Line Ml—MZ vs. F3 335 F1 vs. F3 976 W (P. < 1%). * (P. < 5%). TABLE XV 128 A COMPARISON OF THE WEIGHTED MEAN PRODUCTION OF ALL GENERATIONS OF RED DANISH PROGENY AND OF THE FOUNDATION BREEDS P . Foundation Breeds Red Danish Progeny let Pro— crdlas e d ' d. d. “(*1nt Mean St C.V. Mean St C.V. Over Trait Dev. Dev. Foun- dation Milk 8,536 2,305 27.6 9,116 2,407 22.4 +6.8 Butter— fat 354 90 25.3 372 80 21.4 +5.1 Butter— fat Per— 4.14 0.18 4 1 4.07 0 08 2.0 -1.7 centage 129 In order to assess the probable transmitting value for the production characteristics of the Red Danish sires used in this project, a study of their dams' production records was made. Table VII shows the mean milk, butterfat, and butterfat percent— age of the dams maintained in the Bureau of Dairy Industry herd at Beltsville, Maryland. Repeatability Estimates Table XVII shows the intraherd and all—herds repeatability estimates of milk and butterfat production for the foundation cows entered in the Red Dane project. An inspection of Table XVII re— veals that the repeatability coefficients for milk production for the Holstein. Shorthorn, Jersey, and Mixed breeds were harmonious. They Were slightly different for butterfat, however. The coefficient of repeatability for milk production for the Guernsey breed was con— Siderably in excess of the other breeds. The coefficient for butter— fat InPan-ction for the Guernsey breed was also high in relation to the coefficients of the other breeds. When these data were calculated on an intraherd basis, there Was a drop in the coefficients. The greatest reduction occurred in 1he Holstein breed——a decrease of approximately 45 to 50 percent. 130 TABLE XVI ANALYSIS OF FIRST—CROSS PROGENY TO DETERMINE WHETHER HETEROSIS APPEARED IN THE DATA Number of Dam 8: Mean Prod. Cross Daus. Comparisons of Progeny Milk Production Holstein X Dane 187 9,866 Guernsey X Dane 210 8,742 Shorthorn X Dane 157 8,864 Jersey X Dane 138 8,690 Mixed X Dane 47 8,486 Weighted Avg."K 1,285 9,045 Butterfat Production Holstein X Dane 187 383 Guernsey X Dane 210 375 Shorthorn X Dane 157 353 Jersey X Dane 138 379 Mixed X Dane 47 361 Weighted Avg.* 1,285 372 Butterfat Percentage Holstein X Dane 187 3.88 Guernsey X Dane 210 4.29 Shorthorn X Dane 157 3.97 Jersey X Dane 138 ' 4.36 Mixed X Dane 47 4.26 Weighted Avg.* 1,285 4.10 * This includes all first—cross progeny. TABLE XVI (Continued) P . Mean Prod. of Mean rod Parental % Deviation of F d P t of Red M a Crosses from 011.11 ' aren Dane Parent e n P. Mean Milk Production 9,899 9.577 9,738 1.3 7,586 9,577 8,582 1.9 8, 118 9,577 8,848 0.8 8,212 9,577 8,895 -2.4 8,068 9,577 8,823 —4.1 8,536 9,577 9,057 —0.2 Butterfat Production 373 390 381 0.6 341 390 366 2.4 3 l 7 390 354 0.3 382 390 386 —2.0 349 390 370 -2.5 354 390 372 0.00 Butterfat Percentage 3.76 4.07 3.92 —1.1 4.07 4.28 0.3 4.07 3.97 0.00 4.07 4.36 0.00 4.07 4.20 -1.5 4.07 4.11 —0.5 TABLE XVII REPEATABILITY ESTIMATES OF MILK AND BUTTERFAT PRODUCTION OF THE FOUNDATION BREEDS BRED TO RED DANISH SIRES Repeatability Repeatability of Milk of Butte rfat Breed ___________.___ Intra— All— Intra— All— herd herds herd herds Guernsey 0.53 0.69 0.59 0.67 Holstein-Friesian 0.29 0.52 0.30 0.66 Shorthorn 0.30 0.52 0.33 0.60 Jersey 0.30 0.52 0.35 0.57 Mixed 0.41 0.50 0.34 0.55 All breeds 0.36 0.54 0.37 0.66 133 The Guernsey coefficients were reduced the least——from 15 to 20 percent. The coefficients of the other breeds were in close proximity and between the two extreme situations. The combined repeatability coefficients for milk and butterfat production of all breeds were 0.36 and 0.37, respectively. The coefficients of repeatability for the various Red Danish crosses are shown in Table XVIII. These findings approximate those of the foundation breeds (Table XVII). As in the foundation breeds, there were variations within a cross. The coefficients of repeat— ability for milk production were essentially the same for the three crosses. For the butterfat, however, the first cross had a higher coefficient. The combined intraherd repeatability coefficients for milk and butterfat for all crosses were 0.26 and 0.28—-somewhat lower than that of the foundation herds. The coefficients of repeatability for grading up on an all— herd basis are shown in Table XIX. Each foundation breed and its first~ and second—cross progeny are shown in sequence. There Was conSiderable variation in coefficients within each progeny group_ No exact pattern could be observed from these data; how— ever, thel'e seemed to be a tendency for the coefficients of 134 TABLE XVIII REPEATABILITY ESTIMATES OF MILK AND BUTTERFAT PRODUCTION OF THE RED DANISH CROSSES Repeatability Repeatability of Milk of Butterfat Cross Intra- All— Intra- All—- herd herds herd herds First cross 0.25 0.53 0.24 0.74 Second cross 0.25 0.49 0.21 0.61 Third cross 0.30 0.53 0.40 0.66 All crosses 0.26 0.52 0.28 0.67 135 TABLE XIX REPEATABILITY ESTIMATES OF MILK AND BUTTERFAT PRODUCTION ON ALL HERD ANALYSIS BY FOUNDA— TION BREED AND FIRST TWO CROSSES Foundation Breed Repeatability of Repeatability of and Crosses Milk Production Butterfat Production Shorthorn 0.58 0.55 First cross 0.64 0.56 Second cross 0.88 0.78 Holstein 0.52 0.51 First cross 0.61 0.56 Second cross 0.68 0.71 m 0.64 0.62 FiJII'St cross 0.69 0.58 SeCOnd cross 0.57 0.73 m 0.46 0.62 First cross 0.52 0.49 Seccmd cross 0.52 0.53 136 repeatability of milk and butterfat production of the crosses to be Slightly higher than those of the foundation parent breed. Heritability Estimates In determining the heritability estimates of milk and butter—— fat production, the dam—daughter and paternal half—sibs were di— vided into three groups: (1) foundation—first cross dam—daughter pairs, (2) first—second cross dam—daughter pairs, and (3) second— third cross dam—daughter pairs. The paternal half—sib correlations under each group were of the first—, second—, and third—cross pate mal-half sisters, respectively. Table XX shows the estimates of heritability for milk pro— duction for the various groups by the three methods used in this analysis. The heritability—~estimated by the correlation between half-Bibs-~is higher than either the regression of daughter on dam or correlation between dam and daughter. This is logical, since the correlation between half—sibs usually includes some environ— mental correlation, if any exists. The estimates determined for the heritability of butterfat are shown in Table XXI. As can be Seen, the paternal half—sib estimate is highest for all groups. The In ' . eritabllity estimates for butterfat production by all methods, tend TABLE XX 137 ESTIMATES OF HERITABILITY FOR THE MILK PRODUCTION OF THE THREE GROUPS ON AN ALL—HERD BASIS Intrasire Intrasire Paternal Group Regression Dam—Daus. Half—sib Estimate Corr. Corr. I 0.38 :1: 0.08 0.39 :1: 0.09 0.40 d: 0.21 II 0.43 i 0.12 0.43 i 0.11 0.45 i 0.24 III 0.66 d: 0.20 0.62 :t 0.20 0.70 :I: 0.44 TABLE XXI ESTIMATES OF HERITABILITY FOR THE BUTTERFAT AN ALL—HE RD BASIS PRODUCTION OF THE THREE GROUPS ON Intrasire Intrasire Paternal Group Regression Dam—Dans. Half—sib Estimate Corr. Corr. \ I 0.42 a: 0.10 0.44 e 0.10 0.46 :1: 0.21 H 0.36 4 0.12 0.34 e 0.12 0.54 e 0.24 m 0.67 e 0.19 0.72 :1: 0.21 0.80 :1: 0.45 138 to be slightly higher in groups I and III. The paternal half—sib method gave the largest estimate for butterfat production; this was also true for milk production. The heritability coefficients of all the groups were probably increased due to the intrasire analysis not correcting for the intra— herd variance. In view of the intraherd variance entering into these heritability estimates, an intraherd—intrasire analysis should give a more nearly correct estimate of the heritability of butterfat and milk production. It is felt that these heritability estimates could be expressed 011 an intraherd basis by using a reduction factor obtained from the difference between the all—herd and intraherd repeatability herd Coefficients, since the intraherd analysis of repeatability removes that portion of the correlation due to the records being made in the same herd when all herds are considered. The estimates for the third group are exceedingly high and Probably due to only seventy-three observations being made for twentY’rf—‘Vvo sires. If this be the case, it could be concluded that there a*Ppeared to be no difference between the estimates of herita— b flity of the different groups for milk and butterfat production. ”L 139 Discussion and Conclusion Grading—Up Effect In these data, an attempt has been made to determine the improvement——if any-—made by breeding a succession of purebred Red Danish sires to foundation grade dairy cows. This investigation was designed statistically to analyze the accumulated data produced by this project in order to furnish in— formation for the Red Dane breeders in Iviichigan. Several press articles have appeared, but they were written in a popular style. These were good for promotion of the breed, but the information Presented did not explain the mode of inheritance of various traits 0f the Red Danish breed. From a critical examination of the Tables VI to XV, one can Observe the effects that the Red Danish sires have had on the means of these production characteristics-—both in an average ef- fe°t and a grading—up effect of each foundation breed. Table XV gives a picture of the entire effects that the grading—up project had on milk and butterfat production and butterfat percentage. The Red Dane progeny——one—half, three—fourths, and seven—eighths Red Danish-——increased the mean milk production 580 pounds, or ....: 140 6.3 percent, which is a highly significant increase over the com— bined foundation herd average. The crosses increased the butterfat mean by 18 pounds, or 5.1 percent, which is also statistically sig— nificant at the l—percent level. The mean butterfat percentage was lowered seven—hundredths of 1 percent (1.7 percent), which is non— significant. This increased mean milk and butterfat production was due to the grading up of the Guernsey, Jersey, Shorthorn, and Mixed breeds. The coefficients of variability of all three traits were lower in the Red Danish progeny than in the foundation cows. This indicates that on an over—all basis the crosses were signifi— cantly less variable than their dams (”t" test P < 0.01). The results previously stated show the effect of successive pureb red Red Danish sires on the various foundation breeds. In conclusion, it was found that the average milk production of the various crosses resulting from the Guernsey, Shorthorn, and JerSey foundation breeds showed a significant increase from genera— ti°n to generation (P < 0.05 and 0.0!). The progeny of the Holstein and Red Danish crosses gave a reduced mean milk production; hOWeVer’ statistics were significant at the 5—percent level in only one instance. The means of the progeny of the Brown Swiss and Mixed breed foundation cows were no different than their dams. 141 Table XVI shows that the inheritance of total milk production was very close to the means of the parental breeds for this char— acter. On an average basis, the first cross was 0.2 percent nearer the better parent; however, this cannot be considered as a signifi— cant deviation from the parental mean. Consequently, one can con— clude from these data that no heterosis was evident. The analysis of the butterfat production data for heterosis shows that the means of the first cross progeny were even nearer those of the parental breeds than for milk production. The weighted average of the first cross was precisely that of the weighted average of the parental breeds. Table XVI indicates that the butterfat percentage exhibited an even greater tendency to be half—way between the parental per— centages. An inspection of the Tables VIII to XIII will reveal that this tendency carried through for all generations. These data are in Conformity with those described by Wriedt (1930), in his report of Red Danish and Jersey crosses in Denmark where he found that the F1 generation produced very near the mean of the parental breeds_ This study does not substantiate the data reported by Gilmore (1952), where he estimated heterosis of nearly 15 to 20 Percent in the data reported by Fohrman (1949). Undoubtedly, the differerites between Fohrman's findings and these studies are 142 probably due to the difference in numbers of animals involved. This study represents 1,285 first—cross or crossbred cows, whereas the study cited by Gilmore reported on 28 crossbred cows. In the chapter on the influence of three nonhereditary fac— tors, statistical treatments were employed to determine whether this increase in mean milk and butterfat prediction was due to time trend and/or other environmental attributes. The mean production for the foundation cows and crosses was plotted separately, and no differences in time trends were observed. It should also be considered that each year as many grade cows are being subscribed to the program as there were at the beginning in 1940—41. If im— Provement in environment and general time trends were in effect, the average of the foundation breeds would be as favorably affected as the Red Danish crosses. Repeatability Estimates The intraherd and all-herd correlations for milk and butter— fat Production bet‘.'eer the successive records of foundation breed cows are showr in .'ab1e XVII. This intraherd correlation measures the degree of permanence of milk production as an individual char— act eristic of each cow. When compared with the variation existing 143 in the whole population of records, it expresses to what extent a cow tends to produce the same amount of total milk yield, lactation after lactation. The all—herd correlation tends to make the corre- lation coefficient larger, since when several herds are considered as a single population, the degree of likeness between records of any particular cow is due not only to her own inherent producing ability, but also to the fact that her records were made in the same herd and under similar conditions of feeding and environment. The intraherd correlation more nearly expresses the degree of likeness due to the inherent ability, if the testing environment within the herd is reasonably uniform and removes the likeness between rec- ords due to being in the same herd. ‘ These repeatability estimates varied in the same direction, When comparing the intraherd correlations with the all—herd cor—- relations, as those reported by other workers. For these data the fill—breed single population repeatability coefficients for milk and fat production were 0.54 and 0.66, respectively, while those on the i111:1-aherd basis were 0.36 and 0.37. An inspection of Table XVII reveals that there was quite a variation between breeds for bOth anaiyses. The Guernsey breed was highest in both cases, and the Mixed breed was lowest. The Holstein, Shorthorn, and 144 Jersey breeds showed a very close degree of likeness in the all- herd coefficients. The all—herd coefficients found in this study are not greatly different from those of other findings. Gowen and Gowen (1922) found 365—day lactation records for liolstein—Friesians to be +0.67 for milk. Harris e1 a_l. (1934) found the correlation between but~ terfat records of the same cow to be +0.56 when herds were con~ side red as a single population. Sanders (1930), in his studies of records made by cows belonging to many herds, was the order of +0.73 for total milk yield. Likewise, in a population of Danish cows from many herds, Gaines and Talfrey (1931) found an average cor- relation of +0.50 for yield of fat--corrected milk. This estimate is lower than the preceding ones which also included herd differences. The reduction of the repeatability estimate in Gaines' and Palfrey's data. may be due to the highly selected group of cows they were dealing with, since a cow had to have ten or more records to be included in their study. Dickerson (1940) pointed out that when Gowen and Gowen‘s (1922) estimate of repeatability is expressed on an intraherd basis it becoInes about +0.50 for milk production. Harris gt al. (1934), expressing their findings on an intraherd basis, found that the 145 £11 ‘herd correlation of +0.55 became +0.33 when calculated on an intraherd basis. Plum (1933), from Iowa Cow Testing Association records, calculated an intraherd repeatability of +0.40. Johansson and Hansson (1940) reported an intraherd repeatability of +0.36. Dickerson and Chapman (1940) utilized the intraherd repeatability as a measure of genetic worth of five different kinds of records. The intraherd correlation between age—corrected 305—day 2X but— terfat records, as used in this study, was +0.34. Verna (1945), in a study of one herd——the Iowa State College Holstein herd-—found the estimates of repeatability of the order of +0.50 for milk and +0-43 for butterfat yield. Laben and Herman (1950) found the esti— mates of repeatability for the University of Missouri herd to be +041 for milk and +0.36 for butterfat production. Chandrashaker (1951) and Chai (1951) found the repeatability of butterfat production to be +0.34 and +0.49, respectively, on an intraherd analysis of H0151:9—in—Friesian cattle in Michigan state—owned herds. The esti— mates of repeatability on an intraherd basis found in this study were of the order of +0.36 for milk and +0.37 for butterfat production. These agree very closely with the statistics obtained by numerous 0the 1' investigators . .. 7 r 3 146 An attempt was made to analyze what the grading—up ef— £th . . . . might have had on repeatability estimates of the various crosses from the different foundation breeds. It was found that an analysis of this type was beyond the scope of this study, when done on an intraherd basis. The results as obtained on a single—population study, all herds combined, are set out in Table XIX. An inspection of the data reveals that very little change occurred in the degree 0f repeatability of records of the foundation breeds and their prog— eny- The repeatability estimates of the graded—up progeny tended t0 be slightly greater than those of the foundation breeds. However, these differences are not statistically different. The repeatability estimates for the various crosses of graded- “P Progeny are set forth in Table XIX. It was found that the single— POPU-lation estimates resembled those of the foundation breeds very closely, However, it is peculiar that when the estimates were ex— PresSed on an intraherd basis, the coefficients for the crosses were Somewhat less than those of the foundation breeds. Within the cross'es the intraherd coefficients of repeatability are rather har— monious and more in agreement than those within the foundation breeds. The high values found for the Guernsey breed on an 1“rah-err! basis tend to make the all—breeds estimate high, since F--_——_—— _. . . -'-- 147 the Guernsey breed represented the largest group of foundation QQWS and this would be reflected in the average estimate. The amallness of the estimates of repeatability of the crosses may be due to unsuitable correction factors used to correct for age at freshening. This is substantiated by the findings of Laben and Herman (1950) and Dickerson (1940), who found the repeatability estimates to be of the order of +0.26 and +0.29, respectively, for records uncorrected for age. It is worthy of note here that the correction factors used for the crosses were mixed—breed factors, 81nee factors for this breed have not been calculated as yet. To know the correlation between records made on an intra— herd basis by the same cow in the American Red Danish cattle P°Pl11ation is of importance from the practical standpoint. If the true ability of a cow belonging to these herds is to be estimated as a. basis for some culling or breeding choices, the exact knowledge 0f the repeatability of records in this particular population is needed. The Coefficients of correlation (estimates of repeatability or r) as found in this study supply that information. Of this population, the thirdrcross graded—up Red Danish females are the most important in e81limating a cow‘s real productive ability, since they are the _ -w- _..._ .. _\—’"' . 148 dams of future herd sires to be used in this project. The supply f Purebred sires from the parent Red Danish herd are exhausted aJId this will necessitate the breeders selecting their own sires (fourth—cross males are registerable if certain production require— ments of the female lines are met), and it will be worth their hav- ing the required statistics for computing an estimate of a cow’s real producing ability for the purpose of certain matings designed to produce bulls of greater transmitting ability. It seems reason— a~1C>1e from the data obtained in this study to recommend the use 0f such figures as +0.30 and +0.40, respectively, for milk and butte rfat production for use in estimating each cow's real producing ability in the third—cross population. The equation given by Lush (1949) for estimating a cow's real producing ability is as follows: [herd average] + [(nr) + (1 + [n -— l]r)] x [cow's own average - herd average], where n is the number of lactations in the cow's average, and r the intra— herd repeatability of records of the same cow. While the above formula gives the procedure by which a Cow's real producing ability for any kind of record may be esti— mated. the accuracy of the coefficients of correlation used in the f . ract1cm, nr/[l + (n — 1)r], will determine in part the amount of 149 Q01Hidence which can be placed on such an estimate. The writer {fiels th . at enough numbers were involved in the calculations of the estimates of repeatability for the first and second crosses—~70? and 442, respectively, for them to be accurate. However, for the third .cross, data on only fifty—seven observations were available since the Red Dane project had been in progress for only ten years at the start of this study. It is quite inconceivable to ex— Pect to have large numbers of third—generation females with two or more production records available in such a period of time. Therefore, it seems necessary that a study of this type be repeated With the records of the third—cross females in a short time. Heritability Estimates The estimates of heritability found in these data approached thoSe found by other workers. Lush and Shultz (1936) estimated the heritability of butterfat production to be 0.25, while Johansson and Han380;)n (1940) estimated the genetic portion of the intrabreed vari- ance in single records to be 0.30 to 0.40 for fat percentage. Gowen (1934) estimated one—half of the variation in milk yield was due to hereditary differences. Most of the studies reported in the litera— t are have found the intraherd heritability of differences in milk and 150 butterfat production records to be of the order of 0.3 to 0.4. When the . . estimates of the present study are expressed on an intraherd basis, they are reasonably within that range. The significance of knowing the estimates of heritability for he Red Danish cattle population is for its utilization in the selec— tion differential formula for determining the permanent improve— ment one can expect in the population average each generation. According to Lush (1949), the increase expected per generation should equal the selection differential X standard deviation X heritability. Using this formula, one could estimate how much se— lection need be practiced to attain a specified increase per genera-— “2011 in milk and butterfat production, or how much improvement per gent: ration should be expected with a specified selection rate. In these data, the heritability of butterfat for the second croSS was estimated to be 0.28 on an intraherd basis. The intra- herd Coefficient is derived by multiplying the all—herd coefficient times that portion of correlation due to the records being made in the Same herd when numerous herds are considered. That portion can be determined by observing the percentage reduction of the re— peatability coefficient from an all—herd basis to an intraherd basis. A ssunning that only a 10—percent selection was practiced with the '- ‘.- 151 8eCond cross, the percentage of the population saved would be 90 Is“Cent, giving a selection differential of 0.20, as shown in Lush's selection differential table. The expected increase per generation with this rate of culling can be calculated as follows: 0.ZO(78)(0.28) = 4.4 pounds butterfat, where 0.20 is the selection differential, 78 is the standard deviation of the mean, and 0.28 is the heritability estimate. Since a generation is composed of something like four Years in dairy cattle, this would be an increase of only 1 pound Of butterfat per year. Lush (1949) points out that the actual per— manent improvement in the population average each generation is a fraction of the selection differential, as epistatic differences causes aPPl'oximately one—~half of this gain, and their beneficial efforts are lost in later generations due to recombination. The actual increase found in this data between the first and second cross was only 1 pound (Table VIII) and not 4, as calculated with the selection differential. This indicates that very little culling——-i.f any——had been practiced. For an improvement of 10 pounds of butterfat per generation, it would be necessary for the breeders of cattle in this population to cull 25 to 30 percent of the cows with the lowest prod“ction records per year. 152 Summary A study of the effect of purebred Red Danish sires on milk and butterfat production and repeatability and heritability es— timates——when bred to cows of various dairy breeds—~was made. All records of 270 to 365 days in length were used and calculated to a 305, ZX mature equivalent basis. A lifetime average was calculated for each cow and used as her producing ability for the Comparisons between generations and heritability estimates. The same records were used for the repeatability study, although on a single—lactation basis. A total of 1,358 foundation cows were compared to their graded—up progeny, both by grading up within breeds and as an all— breed average. It was found that the average milk and butterfat Production for the foundation breeds was 8,536 and 354 pounds, respectively, while the averages for three generations of crosses were 9,116 and 372 pounds, respectively, which was a significant increase (P < 0.01). For the graded—up lines within a breed, it was found that no 3‘ - lgnlficant increase or decrease was evident for the graded—up Progeny of the Holstein—Friesian, Brown Swiss, and Mixed breeds, 153 While the production for milk and butterfat of the progeny of the Jex‘i‘ley, Shorthorn, and Guernsey breeds was significantly higher. There was no evidence of nicking or of heterosis among the various breeds and crosses. The estimates of repeatability were made on a single— population basis and on an intraherd basis. The estimates found for the various foundation breeds were similar to those found by numerous other investigators. The weighted mean repeatability for all breeds on a single—population basis was 0.54 and 0.66 for milk and butterfat production, respectively. When calculated on an intraherd basis, the weighted mean repeatability estimates were reduced to 0.36 and 0.37 for milk and butterfat production, re— sPectively. This was a reduction of approximately 30 to 40 percent “3 figure found by'most other investigations. The estimates of repeatability of the crosses reacted slightly differently from those of the foundation breeds. The single—popula— tlon estimates were essentially identical to those of the foundation ""-~ ‘ —‘ . breeds. When expressed on an intraherd basis, however, they re— duced approximately 50 to 55 percent. The reason for this is not a Pparent, although it may be surmised that since the number of r ecords for the crosses was only slightly more than two per animal, ’74 154 the repeatability was lowered because many of the animals may have freshened in a poor condition——resulting in an increased sec- ond—lactation record, although both were used as corrected records. An inspection of the records revealed that inconsistency. \Vhen calculations of the estimates of repeatability were made on a single—population graded—up line basis, there were no Significant differences between the estimates of the foundation breeds and their progeny. Estimates of heritability were found by utilizing the intrasire regression of daughter on dam, intrasire dam and daughter corre- lation and paternal half—sib correlation for the foundation breed to the fil‘st cross (Group I), the first cross to the second cross (Gr°uP II), and the second cross to the third cross (Group III). The estimates found for the heritability of milk production by the intrasire regression estimate for the three groups were 0.38, 0.43, and 0.66, respectively, while those for butterfat productions were 0'42’ 0.36, and 0.67, respectively. There appears to be some en— Vironrmental correlation in these estimates, since the sires were used both-artificially and in bull blocks. Therefore, dam and da’u‘ghter pairs tended to be more alike due to their records being In ad" in the same herd. The heritability estimates for the third 155 group were considerably higher than those of the other two groups. This may have been due to an increased heritability of production from dam to daughter or by sampling error because of only seventy—three dam—daughter pairs distributed between twenty—two sires. An additional study on an intrasire, intraherd basis should be made when more data are available to substantiate these findings with the third group. EVALUATION OF THE PRODUCTION—TRANSMITTING ABILITY OF THE PUREBRED RED DANISH SIRES USED IN MICHIGAN Review of Literature Introduction Dairy cattle are bred primarily for milk production, a character manifested in only the female sex. The males, however, carry the genetic factors detemiining ability to produce milk, and the male parent exerts an influence essentially equal to that of the f€=Iriale upon the productive ability of the female offspring. Due to the relatively large number of offspring he leaves in the herd, the selection of the herd sire is a matter of great consequence. After tWenty years of sire Selection, or the sixth successive sire, 99 percent of the herd inheritance would be from the bulls selected, While only 1 percent remains from the foundation females. The factors used for determining the actual transmitting qualities of dairy sires are generally accepted today as being milk and butterfat production. Type classification plays a somewhat 11111101. role. Both of these factors are the results of a combina- ti . . on of inheritance and environment working together, Wthh results 157 in making the actual transmitting ability of sires very difficult to determine. That this is generally accepted is attested by the great number of various devices proposed by numerous investi- gators in evaluating or predicting the transmitting abilities of dairy sires. The purpose of this study was to determine the sires with Superior transmitting ability as measured by the production of their progeny and dam—daughter comparisons. Methods for Evaluating a Sire's Transmitting Ability Early writings of livestock breeding emphasize that the most 10gical method to evaluate a sire is by the performance of his progeny. Cruickshank and Bakewell rented out what they as- sumed Were good young sires and then took them back if their Progeny proved Worthwhile. Other measures have been constantly Sought_ However, most sires are usually selected by breeders be— fore they have progeny in production by which to evaluate them. Admittedly, the selection of bulls on the basis of their progeny test is preferred, but it d0es not necessarily invalidate the use Of other measures in evaluating an animal. Since a dairy bull 1 Yields no milk, his transmitting inheritance must be assayed by 158 means of his type, his pedigree, or the production facts of his offspring. Over the past thirty years, many suggestions have been made for working out an estimate of the production facts of a sire’s progeny. In general, the idea has been to get a genetic picture by studying his daughters' or sons' daughters' records. Pearl and associates (1919) proposed one of the early sire indices, based on age-corrected milk production and butterfat per— centage records. They suggested the quartile method of arriving at a bull's transmitting ability. This method consisted of com— piling a frequency—distribution table of milk production and butter— fat percentage of dams and daughters. Then the table was divided into quartiles, and the quartiles were designed as follows: (A) the amount of milk or butterfat percentage above the third quartile line, (B) the amount between the median and third quartile line, (C) the amount between the first quartile line and the median, and (D) the amount of milk or butterfat below the first quartile line. In this way, the change in the milk production or butterfat per- centage between any dam and her daughter may be expressed by two letters. For instance a record of the relative milk produc— tion AC means that the dam's milk production is over the third quartile line and the daughter's milk production between the first 159 quartile line and the median. A milk—production record of DB means that the dam's production is below the first quartile line and the daughter's production between the median and third quar- tile line. The authors recommended that in recording a bull with two or more daughters, it seems best to put these pairs on the basis of one hundred. Thus, a bull with two tested daughters out of tested dams, and one dam is-—-inrespect to milk produc- tion-—in class A, her daughter in B, and the pair recorded as AB; the other dam in class C, her daughter in class A, and the pair recorded as CA; the record for the bull will then be 50AB + SOCA on the two pairs. The extension of this method allows the re— cording of any nmnber of pairs. However, it can become rather complicated as can be seen by the evaluation formula suggested for Hood Farm Torono 20th: llAA + SAB + SAC + llBB + llBC + SBC + 5CA + llCB + SCC + 16CD + 5DB + 5DC + 5DD. This method also left unanswered to the breeder the number of pounds of milk and what fat percentage this bull would transmit to his daughters. Yapp (1924) studied the ability of sires to transmit pro— duction capacity to their daughters. The Advanced Registry records of the Holstein, JerSey, and Guernsey breeds were used, 159 quartile line and the median. A milk—production record of DB means that the dam's production is below the first quartile line and the daughter's production between the median and third quar— tile line. The authors recommended that in recording a bull with two or more daughters, it seems best to put these pairs on the basis of one hundred. Thus, a bull with two tested daughters out of tested dams, and one dam is—-inrespect to milk produc- tion-~in class A, her daughter in B, and the pair recorded as AB; the other dam in class C, her daughter in class A, and the pair recorded as CA; the record for the bull will then be 50AB + 50CA on the two pairs. The extension of this method allows the re— cording of any number of pairs. HOWever, it can become rather complicated as can be seen by the evaluation formula suggested for Hood Farm Torono 20th: llAA + 5AB + SAC + llBB + IIBC + SBC + 5CA + llCB + 5CC + 16CD + 5DB + 5DC + 5DD. This method also left unanswered to the breeder the number of pounds of milk and what fat percentage this bull would transmit to his daughters. Yapp (1924) studied the ability of sires to transmit pro— duction capacity to their daughters. The Advanced Registry records of the Holstein, Jersey, and Guernsey breeds were used, 160 and wide differences in the ability of sires to transmit production were found. A formula was proposed to predict a sire's trans— mitting ability from theSe data. It was expressed in 4—percent milk—fat basis by the formula X = 2A — B. Here X = sire’s transmitting ability, A = daughter's average records converted to 4-percent fat basis, and B = dam's record converted to a 4—per- cent fat basis. Yapp also suggested the conversion of these in— dices to percentages of the average of the breed for purposes of comparison. For example, if the breed average was 400 pounds and a bull had a 400—pound index, his transmitting ability would be 100 percent; if his index was 500 pounds, he would be known as a 125-percent bull, meaning his index was 25 percent above breed average. This general type of method, without correction for fat content, has been widely used. Turner (1925) studied Guernsey sires having ten or more Advance Registry daughters. He further extended the study to determine the relation between production of the daughters and their dams. He concluded that the apparent difference between the sire and dam on their daughters' fat production was due to the fundaznental difference in the index of germinal composition of the male and female. He found that for each 100 pounds of fat per 161 year increase in the production of the dams there is a corres— ponding increase of 15 ponds in the daughters. He assumed, therefore, that the dams contribute 15 percent and the sires 85 percent for each 100 pounds of fat produced by the daughters. From these results, he advanced the following formula as the sire‘s potential transmitting ability: X = [Daus. fat production — 0.15 (dam's fat production)] + 0.85. Then Turner (1927a) in his later studies suggested modification of the formula presented in 192.5. He analyzed Jersey sires with Register of Merit daughters and proposed the formula, X = daughter's average fat production — 0.15 dam's fat production, as the best estimate of a sire‘s transmitting ability. . Graves (1926) studied the transmitting ability of twenty— three Holstein—Friesian sires and suggested comparing the dams' and daughters' records and letting the difference (plus or minus) stand for the bull's index. This system has been adopted by the Dairy Division of the United States Department of Agriculture for proving bulls whose daughters have been tested in the Dairy-Herd Improvement Association Testing program. Goodale (1927) derived the Mount Hope Index from a study of Dr. Gowen's crOSsbreeding experiments with cattle at the Maine 162 Station. After considerable study-—which included the records of fifty—eight Jersey Sires, fifty-eight sires of various breeds, and a large series of Guernsey records——Goodale formulated the fol- lowing index as a guide for breeding operations at the Mount Hope farms. This, incidentally, attracted wide attention among other breeders. Briefly the index was formulated as follows. When dams exceeded the daughters, the Sire‘s Milk Index = daus. ave. + 0.429X (daus. ave. — dams ave.). The Sire's Fat Index = daus. ave. + 1.5X (daus. ave. — dams. ave.); when the daughters ex- ceeded the dams, the Sire's Milk Index = daus. ave. — 2.333X (darns ave. - daus. ave.). Sire's Fat Index = daus. ave. — 0.6777X (dams ave. — daus. ave.). Gifford (1930a) made an analysis of the progeny performance 0f Holstein sires in order to determine the transmitting ability of the Sires. The sires were grouped according to their average PI'Ogeny performance records in groups ranging as follows: 500 to 599, 600 to 699, 700 to 799 pounds, and 800 or more pounds b Intel-fat, Coefficients of correlation were calculated for each sub- group_ This grouping had a tendency to hold the sires fairly con— Sta nt, and furnished data which were somewhat less influenced by the - . nFurlllznum entrance requirements than were the data representing 163 the entire population. The weighted—average coefficient of corre— lation obtained was 0.197 and is substantially identical to the re— sults obtained by calculating the partial correlation. Gifford con- cluded that since Holstein—Friesian cows, on the average, only contributed approximately 0.20 of a pound of butterfat for each pound increase in their records to the yearly butterfat record of their daughters above the potential transmitting ability of the sires with which they were mated, it is possible to obtain the value of the sire's transmitting ability where there are relatively large numbers of dam—daughter pairs by means of the following formula, Spt. = d — 0.21), where Spt. is the sire's transmitting ability, (1 the average yearly butterfat records of the daughters, and D is the aVerage yearly production of the dams of the daughters. Gifford (1930b) studied eighteen Guernsey sires, selected at ran— dom, that had twenty or more daughters out of tested dams. He also Calculated the Mount Hope Index (Goodale, 1927) and the dif— ference in production as used by the United States Department of Agriculture (Pearl and associates, 1919; Graves, 1926)» and com- Pared the indices with all the daughter's average production of a Sire, and concluded that the daughter's average was of more value 164 in predicting records of daughters than the sire indices presented by the other investigators. Wright (1931) reviewed the formulas suggested by Yapp (1924) and Goodale (1927) for estimating the transmitting ability of dairy sires. He concluded that Goodale's formula was biased because of being formulated from the performance of crosses of two breeds as abstract as the Jersey and Aberdeen Angus. How— ever, in taking an average value of the two formulas suggested by Goodale and comparing it with the ones proposed by Yapp, he found the two to differ only slightly. Wright was critical of Yapp's formula. in that it did not offer a measure for taking into account the number of daughter—dam pairs available in the comparison. Wright suggested a formula based on the genetic theory of multiple— factor inheritance, giving weight to the breed average, where the number of daughter—dam comparisons is small and increasing Weight to the daughter—dam comparisons as the number of these comparisons increases. He prOposed the following formula: 2 = [2/(n + 2)][A + n/(n + 2)][20 — D — A], where n = the number of d:‘lug'l'ZI-ter--dam comparisons, A = the breed or herd average in production 0 = daughters' average production, D = the dams' } p \ V 1 , t 1 l ” l ‘ \ l w . ‘ 5 i L l t l > . H . l . ‘ V t , 165 average production, and Z) = the bull's transmitting ability or index value. Ward and Campbell (1940) reported on .the "sire-survey work" in New Zealand. Under the program, when the first daugh— ters of a sire have completed a testing year, a "preliminary" survey of the sire's performance is issued to the breeder. The survey is continued from year to year. At the end of the second testing year an "intermediate" survey is issued. This contains the performance of the second records of the sire's daughters and any additional first records, second records, and third records 0f daughters would be included in the survey; this would be issued in the: form of a "final" survey of the sire. Age—correction fac— tors (New Zealand data) are used only for two—year and three—year recoll'ds. The New Zealand workers preSent an evaluation of forty- one Sires with "final" sire surveys. The general use of a single index by which to express the breeding value of a bull is not likely to be Very accurate in practice, but the average production of all daughters is a better approximation to the bull’s probable breed— ing Value than the intermediate parent index. They also suggest that I“any of the Serious practical limitations ascribed to sire indie . . -. . . e3 are more academic than practical in their implications 166 and that the practical breeder or dairy farmer can make use of sire—Survey work. Rice (1944) and Rice and Andrews (1951) reported that because of its simplicity and approximate accuracy, the “Equal—Parent” index, also known variously as the "intermediate," ”American," or "Yapp— Hansson," is being widely used today. The "Equal—Parent" index is generally calculated by finding the differences between the average amount of milk and butterfat test of a group of cows and the similar averages for their daughters by a given bull. For both amount 0f milk and butterfat test, if the daughters' average ex— ceeds the dams' average, twice the difference is added to the daughter's average to secure the bull's "Equal—Parent" index; if the daughter's average falls below that of the dam's, twice the diffel‘ence is subtracted from the daughter's average to secure the index. This method, therefore, always places the daughters intermediate between the dam's actual average and the bull‘s ll Equal—Parent" index. This can be resolved into the formula EP~ ‘ 2d — m, where EP equals bull’s index, d equals daughter's average, and m equals dam's average. Rice (1944) examined the bull—index problem and proposed a . . new Index called the Regression Index. It is based on the findings that the regression of daughters' records on those of groups of dams was 0.5. The regression index can be found by getting the difference between a sire's daughter's actual and "normally expected" production and adding this difference to the breed average. Rice suggested that if normal tables of expecta— tion were not available, the normal expectation for any group of daughters could be ascertained by adding the respective breed averages in milk and test to the dam's average milk and test and dividing by Z. This is done on the basis that daughters regress 0.5 toward breed average from their dam's average. Similarly, an Optional manner and the easiest way of finding a bull's Re- greSSion Index would be to find his EP index in the usual fashion, twice daughters' averages minus the dams, then add the respective breed average and divide by two. Allen (1944) used the data compiled by the nationwide sur— Vey of dairy sires proved in dairy—herd improvement associations by the Bureau of Dairy Industry which made available 5,886 proved Sires 0f the five major dairy breeds. The data for each breed Were COnsidered separately. He assembled theSe data according to the dams to which the individual sires were mated. Milk produc- ti on, fat production, and fat percentage were considered separately. 168 From these data~-for each breed and for each of the three per— formance factors-—the average level of performance of the sires' progeny was estimated in relation to the average level of the dams to which they Were mated. These findings were set up in graphical form, to be used as a guide in evaluating progeny—tested bulls and in predicting probable production of future daughters. Dickey and Labarthe (1945) have proposed a system to evaluate the probable transmitting abilities of young bulls by an analysis of their pedigrees. Data for this study were obtained from USDA Miscellaneous Publications No. 487, June, 1942. Each HOIStein—Friesian proven sire listed in this publication was used and éalled a "son." Then the sire and dam of each of these "sonslv were determined. If the sire was proven and the dam had a. record, the "son" and his pedigree was uSed. The investi- gators succeeded in obtaining 241 such pedigrees for this study. The formula used for predicting records of "sons' " daughters by means of the regular pedigree value was PR = (C + D) + 2, Where PR is the predicted records, C is the son's pedigree value found by adding Equal-Parent Index of sire to son's dam's record and dividing by Z. and D is the son's mate's records. The cor— 1.el‘a‘tj-Oil'i between the predicted records and the actual records of 169 the son's daughters were 0.565 for milk production, 0.531 for butte rfat yield, and 0.630 for butterfat test. At the time the study by Dickey and Labarthe was in prog- ress, Rice (1944) announced his regression index for proving dairy sires, and the authors applied a regressed pedigree value to these data for predicting son's daughters' production. The procedure was the same except for finding the son's pedigree value equals C = (A + B) + 2, where A is the Regression Sire Index (equal—— parent index + breed average) 0.5, and B is the dam's transmis- sion Index (her record + breed average) 0.5. By this method the C°1‘I‘elations between predicted and actual records were slightly higher than by the regular pedigree value method. Also, when the Values were plotted to a straight—line regression the regressed pedigree values more nearly fitted a straight line or approached one fol? milk and butterfat yield. This led the authors to con— Clude 1Chat the regressed pedigree value is superior to the regular Pedigree value in predicting the transmission of milk production and butterfat production; however, the two methods are about equal in Predicting the transmission of butterfat test. They also sug- gested that perhaps by regressing the "son's mates" production with the average breed production, or by applying correction factors .._—-= 170 to the "son‘s" pedigree at varying levels of production, some re— finement c3n be added to the prediction_of milk and butterfat pro- duction and butterfat test. Washbon (1947, 1948, 1950, 1951, 1952), in a series of publications, has undertaken the evaluation of a sire's transmitting ability through the performance of_his sons. He has studied all bulls proven by the Division of Dairy Herd Improvement, Bureau Of Dairy Industry, that have five or more proven sons. Bulls studied in the Holstein breed include those with official Holstein association proofs. The bulls were all proven by the method of GraVes (1926). He has constructed tables which evaluate all proven sires by the proof level of their sons. Eldridge (1948) presented a "Multiple Regression Method" of Sire selection. He studied the proof levels of 1,451 Holstein Sires proved in the State of New York and listed in the Animal Husbandry Extension files of Cornell University. A method was Cleveloped for selecting young, unproved dairy bulls that predicted the Production of the daughters of such sires. The formula, {VA ‘ 29-6 + 0.75X + 0.03): + 0.01): + 0.34X4 _ 0.21x 2 3 5, developed by E ldridge, gave a predicted production value which was correlated 0. .. 70 With the actual production. Another formula presented by 170 to the "son's" pedigree at varying levels of production, some re— finement can be added to the predictionpf milk and butterfat pro— duction and butterfat test. Washbon (1947, 1948, 1950, 1951, 1952), in a series of publications, has undertaken the evaluation of a sire's transmitting ability through the performance of.hi_s sons. He has studied all bulls proven by the Division of Dairy Herd Improvement, Bureau of Dairy Industry, that have five or more proven sons. Bulls studied in the Holstein breed include those with official Holstein association proofs. The bulls were all proven by the method of GraVes (1926). He has constructed tables which evaluate all proven sires by the proof level of their sons. Eldridge (1948) presented a ”Multiple Regression Method" Of Sire selection. He studied the proof levels of 1,451 Holstein sires Proved in the State of New York and listed in the Animal Hquand-ry Extension files of Cornell University. A method was developed for selecting young, unproved dairy bulls that predicted the Production of the daughters of such sires. The formula, TA = 5, developed by z 9'6 + 0.75}: + 0.03XZ + 0.01:»:3 + 0.34x4 — 0.21x Eld ridge, gave a predicted production value which was correlated 0.70 " With the actual production. Another formula presented by 171 1 Eldridge, i' = 0.1 + 0.75):1 + o.o1x + 0.23x _ o.zzx + 0.24x , B 3 4 5 6' gave predicted values correlated 0.69 with actual production. ?A is the predicted average mature equivalent production of the daughters of the bull, when X1 is the average production of the cows to which this bull I "is to be" or "was" bred. X2 is the average production of the maternal half—sisters of the bull. X3 is the production of the dam of the bull based on the aver— age of the maximum number of records available. X4 is the average production of the paternal half—sisters of the bull, and X5 is the average production of the dams of the paternal half— Sisters. A YB is the predicted average mature equivalent production of the daughters of the bull, when X 2 is replaced by X which is the average production of the 6, daughters of the maternal grandsire of the bull. EdWai-ds (1932) analyzed production records of Jerseys in En gland ‘30 estimate the number of daughters needed to give the best esti”5118.13. He also applied five of the indices discussed 172 previously——Pearl and associates (1919), Yapp (1924), Wright (1931), Goodale (1927), and Gifford (1930b)——to these data and concluded that the daughters' average production (Gifford, 1930b) was the better for determining a sire's transmitting ability than any of the others. However, he cautioned that all data must be studied closely and that these findings might well not apply to other data. His findings bear out those of Gifford (1930b). Rice (1944) sets up a criterion that a bull index should meet, as follows: sound from a genetic viewpoint, easily understood, cal— culated in terms of breed average, and comparable in variability t° grOU-ps of animals rather than to individuals. According to Rice, a bull i“Clem: should also perform the following: rank bulls in their proper Order, provide a definite measure of a bull's transmitting perform“lance, provide a means for predicting future daughters' produCtiOn, and provide as accurate a means as possible for eval— uating Pedigrees. He measured the Equal Parent Index and the Regre'SSion Index against the above-named criteria and concluded the Regression Index more nearly fitted the expectations. Lush and Nelson (1943) evaluated the methods available for measuring transmitting ability in sires and cows and the biased err Ors which enter into such evaluations. They suggested that in 173 most data the biased errors will be large enough that the random errors are already reduced to secondary importance by the time there are as many as five daughters. Further increases in the number of daughters raises the accuracy of the estimate of trans— mitting ability only a little unless those biased errors are reduced by other methods. Lush and Nelson advanced four principal sources Of biased errors: (1) The general environment under which group of daughters made their records may vary from sire to sire. (2) The general environment of the daughters of a sire may not have been the same as the general environment of his mates. (3) The average genetic merit of the mates may vary from sire to sire. (4) Selection of the mates may have been more intense for some Sires than for others, thus making the average record of the mates farther above their average transmitting ability in some cases than in °thers. In applying these biased errors to the most widely used measures of transmitting ability, (.k) the average of the daughters, (B) the increase of daughters over dams, (C) the various indices such as equal—parent, Mount Hope, Yapp—Hansson, etc., using the sum of A and B, the writers concluded that these measures are affected as follows by the biased errors. A is affected most and 174 B least by error (1). B is affected most and A not at all by (2). A and B are affected in opposite directions by (3) while C is un~ affected. B is affected most by (4) while A is not affected. They concluded that error (1) is generally the most serious source of error although (2) can be in some cases. Selection of mates (4) can only rarely be of major importance since opportunity to vary widely the intensity of selection of mates will not often occur. Measure C is believed to be more widely useful than the other two but, since the measures are not equally vulnerable to all kinds of errorS, there may be some sets of data in which measures A or B Would be more accurate. If there are data in which it is known that errors (2), (3), and (4) were very small or zero, while error (I) was large, measure B would be more accurate for analyzing 1:heSe data. However, if the reverse is true, error (1) is small and errOrs (2), (3), and (4) are large, A will be more accurate. They Suggest that where some or all of the dams are untested, measures B or C can be used to good advantage by using in place of the misSing records the average production of the contemporaries of the daughters in the same herd. This helps correct for errors (1) and (2). Minimum Number of Progeny for Sire Evaluation The question of how many progeny should be considered as a sufficient number to be used to measure the relative breeding value of a sire has been a prevalent one. It is evident that, as the number of progeny considered in a progeny test increases, the resulting predictions are more accurate. However, there must be a point at which an increase in numbers is of little value in in— creasing prediction accuracy and likewise there must be a point at which a decrease in numbers will cause an increase in inac— curacy. The Bureau of Dairy industry has adopted the number of progeny to be five when using daughters of a sire in a ”sire prooffl! Washbon (1947), in his “Sire Directory," used five or more a-8 the number of sons for evaluating a sire. Several investi— gators haVe undertaken to determine statistically the optimum num— be . . . r 0f Progeny to use in Sire evaluations. % Davidson (1925) studied the Register of Merit of The Amer— i can Jersey Cattle Club to determine the number of tested daugh— ters of a sire whose average production could be used as a relative 176 measure of a sire's breeding value. He selected all Jersey sires having at least fifteen tested daughters. From a preliminary study of the data Davidson found that it may be assumed that the varia— bility in the production of the first fifteen daughters of a Jersey sire is representative of the variability in the productions of any large number of his tested daughters since the mathematical con— stants used to measure the variability in the production were many times their probable error. He also used the same assumption in using fifteen daughters as the basis with which any smaller number of tested daughters may be compared. In order to determine the smallest number of a sire's first tested daughters whose average PrOdUCtions would closely approximate any large number of his tested daughters, Davidson set up a classification listing the daugh— ters by chronological order as they appeared in the Register of Merit, Then a grouping of these was made: the first daughter made the first group, the first and second daughters made up the second group, and so forth, until the first fifteen daughters made up the fifteenth group. The coefficients of correlation were calculated between each group with the fifteenth group of daughters and were fitted to a curve. The curve rose very rapidly at first and then gradually flattened out into an almost straight line. The point at 177 which this flattening in the curve began to be pronounced lies in the region of the sixth coefficient of correlation, which represents the sixth daughter—groups. The coefficients calculated from the fitted curve increased from +0.526, the coefficient of the sixth ag— gregate, to +1.000 for the coefficient of the fifteenth aggregate. Approximately three—quarters of the total increase in the correla-- tions from the first to the fifteenth aggregate is reached by the sixth aggregate. thus showing that after the sixth aggregate, addi— tional daughters influenced the correlations only slightly. Davidson the refore concluded that on the average, the mean milk yields of the first six tested daughters of the 133 sires studied, are a very close approximation to the mean milk yields of their first fifteen tested daughters. Lush (1931) investigated the minimum number of progeny needed to make a rather reliable prediction of a sire's future daugh- ters_ He set up a general formula for the correlation between the average record of the daughters and genotype of the sire for factors a‘ffectilig milk production. Lush suggests that if we let "5" rep—— resent the path coefficient from the sire's genotype to the daughter's record, and if we let ”e" represent the path coefficient from the he rd nmanagement or common environment (including also the average ...|.. 7. ‘4.an tow-@VP h. ll o a. 178 relationship of the dams to each other) to the daughter's record, then it is evident that the primary observed correlation between the records of half sisters equals e2 + 52. The value of 3 would be 0.50 in a random—bred population where there was no dominance and Where milk and fat production were entirely determined by heredity. The value of e2 is 0.15 as determined from Advanced Registry data. Expected correlation between sire's genotype and the ave rage record of various numbers of daughters at selected values of s and e2 were calculated. In the calculations, 3 had been figured at four different values; namely, 0.30, 0.40, 0.50, and 0.60, Whereas e2 was figured at 0.00, 0.10, 0.20, 0.30, and 0.40. Lush concluded from the calculated curves using different values of s and e2 that at low values of s and e we may gain a very notice— able 1Increase in the reliability of the progeny test by increasing the 1lu-Itnber of daughters even past eight or ten. At the other ex— treme, with high values of e and 5, there is little accuracy to be gained by increasing the number of daughters even past three or four- For what appears to be the most probable values for s and e, °n1v a. little increase in accuracy is to be gained by including more than five daughters in the progeny test. Of course one would L; 178 relationship of the dams to each other) to the daughter's record, then it is evident that the primary observed correlation between the records of half sisters equals e2 + 52. The value of 5 would be 0.50 in a random—bred population where there was no dominance and where milk and fat production were entirely determined by heredity. The value of e2 is 0.15 as determined from Advanced RegiStI'y data. Expected correlation between sire's genotype and the 3V8 rage record of various numbers of daughters at selected values of s and e2 were calculated. In the calculations, 5 had been figured at four different values; namely, 0.30, 0.40, 0.50, and 0.60, Whereas e2 was figured at 0.00, 0.10, 0.20, 0.30, and 0.40. Lush conchl(led from the calculated curves using different values of s and 22 that at low values of s and e we may gain a very notice— able itlcrease in the reliability of the progeny test by increasing the I“ltl‘iber of daughters even past eight or ten. At the other ex— treme’ with high values of e and 3, there is little accuracy to be gained by increasing the number of daughters even past three or four. For what appears to be the most probable values for s and e, 0111? a little increase in accuracy is to be gained by including m ore than five daughters in the progeny test. Of course one would 179 want to base his estimate on all that are available, no matter how many there may be. Edwards (1932) employed the formula, V = [Sum (for all groups) of sum of squares of deviations from group mean] + [Total ._ 2 daughters minus number of groups], or V = [2(x — x) for all groups] + [T daus ~ N groups], to calculate the variance of one daughter, in order to obtain an indication of the minimum number of daughters necessary to prove a sire. By this method Edwards calculated the 0' of the milk yield of a single daughter and the a" for the mean of an increasing number of daughters up to fifteen by the corresponding root of the number. He found that for two sires with one daughter each, the difference—40 be significant—— between milk yields had to be 5,521.5 pounds; with 15 daughters each, it had to be 1,425.6 pounds; with 52 daughters, 770 pounds. The anthor concluded that, since a breeder desires to know with a minimum of daughters the value of a sire, six daughters are nec- essary to evaluate the ability of a sire. With six daughters the difference to be significant between the averages of the milk yield Of daughters of sires must be 2,469.3 pounds. Washbon (1949) investigated the possibility that "D. H. I. A. Proved Sons" of a sire might be used to evaluate the inheritance of a sire or a sire family. -One hundred and seventy—four Hol— stein sires with more than fifteen proved sons were studied to de—— termine the least number of proved sons necessary to most accu— rately estimate the performance of those proved later. Highly sig— nificant correlations (r = 0.58) indicated that the average butterfat PrOduction of the daughters of the first three proved sons was nearly as accurate as data on more sons in estimating the average bUtterfat production of the daughters of the next three, five, or ten Proved sons of a sire. Similarly, the significant correlations (r = 0.33) for the son's daughters' increase or decrease in but— terfat production from their dams indicated that data on the first three or four proved sons were nearly as accurate as data on a larger number in predicting what might be expected from the next three, five, or ten proved sons in this respect. For the percent— age °f Sons showing plus proofs, the correlations were significant When the first four, five, and six 'sons were compared with the “3’“ ten (r = 0.25). 181 Investigational P rocedure Numerous investigators have shown that the production of the Progeny of a sire and the dam—daughter comparisons within each sire are as good a measure of a bull's transmitting ability as any of the available methods. These two methods were used in this study. The sires were grouped according to three sire—family groups, since all bulls used as sires are either sons or grandsons of the three imported bulls (one imported in dam). The mean pro— duction, standard deviation, and coefficients of variability for the Progeny of each sire were calculated using the same procedure as Pre viously described. All data pertinent to this part of the study were obtained f1301'!) the IBM Card No. 6, ”Heritability and Sire Evaluation Card" (Fig. 4). A series of sortings, calculations, and tabulations was Performed with the final products printed on the usual folded sheet Of Pipe r. Inve stiggtional Re 3 ults An inspection of the pedigrees revealed that the imported sires D501, D502, and D505 had thirty-four, thirty—three. and 182 seventeen sons and grandsons, respectively, used in this project. In order to determine whether differences between sire families estted, the sons and grandsons were grouped according to the three imported bulls. Table XXII shows the statistics obtained for each sire within a sire family. In analyzing these bull fam— ilies, it is seen that the production of the daughters was essen— tially the same for families 50] and 505. The progeny means in sire family 502 were on the average 10 pounds below the level of the progeny of the sires in the other two families. It could be Possible that by random chance the level of production of the mates Was lower for that family; however, this is not probable, since the Sires were moved each year to a new herd and some from each group were used artificially; therefore, the chance of being mated ‘10 mates of different levels was unlikely. However, when rated according to the sire's transmitting ability as calculated from these data, there appeared to be no apparent difference in the sire fam— flies, It is interesting to note that each family had one or two sires that have a Substantial minus proof level. On the whole, howeVer, the proof levels were good since 81 percent of the sires increaSed production over their dams. TABLE XXII PERFORMANCE OF RED DANISH SIRES BY FAMILY GROUPS No. . I d- P f Slre :1?er No. Avg. Std. Coef. Dam— If”: Sire's NO- ,g Daus. Fat Dev. Var. Daus. 1 T.A. Sire , ference Pairs Sire Family 501 1507 00.8 34 338 67 19.3 15 ~ 8 340 1561 00.8 45 376 65 17.5 30 +12 388 1562 03.1 25 304 59 19.4 13 —49 262 1569 00.4 17 326 79 24.2 13 —36 282 1572 00.8 21 377 72 19.4 17 +40 464 1 575 00.8 60 396 56 14.1 50 +35 428 1 578 00.8 27 373 68 18.2 16 +18 444 1 6031 37.5 46 376 69 18.4 24 +31 440 1607 00.0 46 388 89 20.9 32 +16 416 35.7 369 74 20.0 23 +15 405 Sire Family 502 29 332 53 16.0 15 +32 383 45 375 64 17.0 32 +25 429 103 356 67 18.8 74 +22 407 62 350 68 19.4 35 +34 411 184 7 TABLE xxn (Continued) Sire 1:31:32? No. Avg. Std. Coef. D11;— pg: Sire’s3 No Sire Daus. Fat Dev. Var. Daus. ference T.A. Pa1rs 1559 03.1 72 373 74 19.8 39 +32 435 1560 00.0 7 351 55 15.7 1567 00.0 19 384 76 19.8 15 + 8 419 1568 13.3 27 375 65 17.3 22 +30 438 15821 14.3 28 320 95 29.7 21 —48 246 1604 13.3 24 361 59 16.5 12 +73 504 1611 00.0 40 350 68 19.4 25 + 2 365 Avg.2 41.5 358 74 20.7 29 +20 402 Sire Family 505 1563 07.8 17 372 61 16.4 10 +11 372 1564 07.8 21 351 57 16.2 15 + 4 350 1566 00.6 46 373 74 19.3 24 +35 441 15701 00.2 35 413 67 16.2 31 +47 499 15711 00.4 46 347 63 18.1 30 — 4 339 15731 00.2 34 368 52 14.1 26 +17 403 15741 01.0 11 359 85 23.6 7 —47 285 185 TABLE XXII (Continued) No. 1 d— f Sire 2:"; No. Avg. Std. Coef. Dam— Fri": Sire's3 N0. g Daus. Fat Dev. Var. Daus. T.A. Sire . ference Pairs 15771 00.0 7 395 64 16.2 4 —10 365 15791 00.0 14 342 56 16.3 11 +42 426 15801 14.3 12 375 62 16.5 11 +20 412 16021 07.8 26 352 36 10.2 14 +13 382 11ng 24.5 368 69 18.8 16.6 +18 404 Grandsons of imported sires. Weighted average based on dam—daughter pairs. Sire's transmitting ability determined by adding twice the daughter's difference to the dam's average. This is based on results obtained from the study on grading up. 186 To determine the effect of the inbreeding of a sire on the variability of his get, a study of Table XXII was made. There were seven sires that were inbred more than 7 percent, or an average of 14.5 percent. In studying the standard deviations and coefficients of variability, it appeared that the standard deviations were less for the gets of the inbred sires than the standard devia— tions of comparable means for the gets of the noninbred sires. However, when a weighted analysis of the data was made by pool- ing the total butterfat production and sums of squares for the inbred and noninbred groups, a different conclusion was made (Table XXIII). These results show no apparent differences between the variance of the gets of noninbred sires as compared to those from inbred sires. Dis cus 3 ion and Conclus ions An attempt was made in this study to separate the Red Danish sires into sire families and analyze their transmitting ability by daughter averages and dam-daughter comparisons. As was expected, three sire families were identified, because all the Red Danish sires used in Michigan originated from the three im- ported sires. To date, there has been only one importation; m_ il’ fi ,- 187 TM3LE XXIII RELATION OF VARIANCE OF PROGENY TO INBREEDING OF THE SIRE Avg. G I.B.C. No. Avg. Std. Coef. ‘ roup of Daus. Fat Dev. Var. 1 Sire I Noninbred 0.0 686 367 71 19.3 Inbred 14.5 173 370 73 19.7 consequently, the pedigrees of all Red Danish animals will trace baCk to the animals in this importation. It was not expected, however, to find such close conformity °f the progeny averages and dam—daughter comparisons for the three families. On inspection of Table XXII, one sees quite a large Variation between sires within a. family, but on a family av— erage the three are rather close. One of the families can perhaps be Considered slightly inferior to the other two. In analyzing the reaSOn for the apparent 10 pounds difference, it appears that the Sire , s of family 502 were mated to cows of a lower mean produc— tion_ The proof level difference between dam—daughter comparisons 1 38 was +20 pounds, which is slightly more than found in the other fam— ilies. This conforms with the findings of Table XVI, where it was observed that the production of the progeny was essentially that of the mean of the two parents. So if by random chance, the sires of family 502 were mated to cows with a lower mean production than those of sire families 501 and 505, i' is very conceivable that the results such as observed can be expected. To analyze this difference more thoroughly, the average production of the mates of each group must be calculated to determine if the averages were different. i In analyzing the transmitting ability of these Red Danish sires, the effects of selection of the foundation cows must be con~ side red, Prior to entering the program of breeding to Red Danish sires, it must be assumed that normal selection or culling was PraCtiCed by the breeder. According to Lush (1949), there is ap— Pr°ximately a ZS—percent turnover of herd numbers, with over half 0f theSe for reasons of low production. Gilmore (1952) stated that for each 10 percent of the low—producing daughters of a sire that are culled, the average production of the remaining daughters in— creases 12 pounds. When a dairyman entered the Red Danish pro— :1 gr m, he agreed to raise all Red Danish progeny, milk them at 189 least one lactation, and to remove all foundation cows as rapidly as possible. 'lhis regulation placed the Red Danish sires under the handicap of not having their daughters undergo the normal se— lection for production which occurs in all other herds. Also, the first year that Red Danish heifers came into milk production in a herd, some of the foundation cows were culled in order to make room for the freshening Danish heifers. This, in turn, increased the selection of the foundation cows to an even greater degree. After several such selections of the foundation cows, only those with the highest production records remained. Certainly the Danish sires bred to these highly selected mates were at a disadvantage, since this same type of selection was not practiced with their daugh— ters. An added impetus to milk everything, so to speak, was in- duced by the formation of the open herd book of the American Red Danish Cattle Association. The rules allow for a third—cross fe— male to be registerable, provided she makes a production record (no required amount), and provided all the females including the foundation cow in the graded—up line be production tested. The investigator feels that it is essential for one to know these exist— ing conditions to properly evaluate and criticize the production traits of the progeny of the Red Danish sires. There is no direct 190 method of measuring the difference but it is felt that the Red Danish sires were proved under a minimum handicap of 20 to 25 pounds of butterfat when applying the observations Inade by Gilmore (1952) to these data. Summary Thirty—one sires were evaluated for their transmitting ability by the mean butterfat production of their daughters and a dam—daughter comparison. They were grouped into three sire families for analysis. 'i'hc average production of the progeny from the three groups was very similar; however, the progeny for family 502 was 10 pounds lower than the other two families. When comparing the families according to their average transmitting ability, it was found that there was a still greater uniformity between the families. fiitlzin families there was found a great variation between the average production of the different sire's progeny as well as their transzrzitting ability, Eighty—one percent of the sires had plus proofs. Seven of the thirty—one sires were inbred an average of 14.5 percent. An analysis was made to determine the effect of inbreeding a sire on the variation of his progeny. 'I‘he daughters 191 of the inbred sires were compared with those of the noninbred sires and no appreciable difference was found in the variability of the two groups. THE EFFECTS OF MILD INBREEDING ON MILK AND BUTTERFAT PRODUCTION AND THE OCCURRENCE OF DEFECTS Review of Literature Introduction In the development of many of our breeds of livestock, in— breeding has played an important role. Bakewell successfully used this method in his breeding operations with sheep and Short—- horn cattle. Yoder and Lush (1937) reported that the Brown Swiss breed in the United States is descended entirely from 129 cows and 21 bulls which were imported to this country. Over half of the random pedigree lines of the Shorthorn breed go to one bull, Favourite, and, in American Rambouillet pedigrees, about 45 per— cent of the lines traced back at random end in sheep from the von Homeyer flock in Germany. Even though inbreeding has successfully molded various breeds, it has found little favor in practical breeding programs. Its deleterious effects have been long known and well recognized. However, it is quite generally recognized that superior stock can 193 be successfully inbred. In recent years and at the present time, inbreeding projects are being conducted by a number of experiment stations to obtain a more conclusive answer on how to derive greater benefit from this system of mating. The main purpose of this study was to determine the effects of mild inbreeding on the yield of milk and butterfat production and ‘ the occurrence of lethal defects. Inbreeding of Laboratory and Farm Animals The studies by King (1918, 1919) on rats reopened the pos— sibilities of utilizing inbreeding effectively in animal improvement. At the end of twenty—five consecutive generations of full—brother full—sister matings, the rats were superior in size and fertility to the stock of outbreds carried as a check. However, selection for those two characteristics was very severe. Wright (1922), reporting on the inbreeding of guinea pigs by the United States Department of Agriculture, found that on the average, at the close 0f twenty generations of full—brother times full—sister mating of guinea pigs, there had been a decline in all elements constituting Vi801'; however, there were several families which had not suffered any Very apparent degeneration. Crosses between different inbred 1 1 L ,5 - l 194 families resulted in a marked improvement over both parental stocks in every respect. Morris, Palmer, and Kennedy (1933) developed two strains of rats from a common foundation, which, in the ninth generation, through continued full—brother times full—sister matings accompanied by careful selection, differed markedly in efficiency of food utili— zation. The low line was 40 percent less efficient and more variable than the high line. Waters and Lambert (1936) inbred White Leg— horn fowl over a ten—year period, during which time they developed six inbred families, with coefficients of inbreeding ranging from 41 to 82 percent. No general decline in fertility and egg production occurred, and the more intensely inbred birds matured sexually sixteen days earlier. On the other hand, Winters and associates (1943, 1948), in appraising the early inbreeding experiments with swine, chickens, and laboratory animals, and giving what is probably the best and latest review in this field, found a. general negative trend. They concluded that these negative results found with farm animals were due to small numbers, thus preventing rigid culling; performance Was not given enough attention; and there was too much adherence t° experimental patterns to give the best results. Willham and L. 195 Craft (1939) and the 1936 Yearbook of Agriculture presented good reviews of the work with farm animals in this field. Inbreeding of Dairy Cattle Nearly all of the data available on the inbreeding of dairy cattle have come from the experiments conducted by the Bureau of Dairy Industry at its Beltsville station and by the New Jersey and California Experiment Stations. More recent inbreeding experi— ments have been inaugurated at several other stations. Some other data have been made available by several workers analyzing the ef— fects of inbreeding as it occurred among private herds——not as a planned experiment, but as a system of mating in trying to fix a definite trait. Production Plum (1933) studied the production records of 183 cows of a. private herd where considerable inbreeding had been practiced. He found a significant negative correlation within groups of daughters by the same sire which indicated a tendency for the more inbred daughters to be the poorer producers. A positive correlation between sires indicated that as a whole the sires with the inbred daughters m 196 also had daughters whose production was above average. Plum concluded that inbreeding in general lowers production. Bartlett 91: a_l. (1939) observed somewhat similar results in analyzing the records of the New Jersey experiment. However, in this case in— breeding had to be considerably high (0.25) in order for it to have a negative effect on production. F2 cows, 25 percent inbred, pro—— duced 2.7 percent less milk and 8.5 less butterfat than did their dams, whereas F cows, 60 percent inbred, produced 31 percent 6 less milk and 38 percent less butterfat than did F1 cows. Asdell (1945) studied the breeding systems used to produce the highest yielding cows of the Guernsey and Holstein—Friesian significance in the production of cattle with the highest milk or butterfat yield. However, he deduced that strong inbreeding was detrimental in preventing the expression of genes for high yield, since the best results were obtained in one Holstein—Friesian herd When the coefficient of inbreeding was reduced. Woodward and Friesians, found results that compare with those of Bartlett e_t al. (1939). Inbreeding did not have much effect, if any, until the co— efficient of 20 percent was reached. Cows inbred 25 to 29 percent dairy breeds. He observed that the system of breeding was of little Graves (1946), studying the effects of inbreeding on grade Holstein— 197 produced 10 percent less milk and 9 percent less fat than did cows inbred less than 20 percent. Cows inbred 50 percent or more pro- duced 20 percent less milk and 25 percent less butterfat than did cows inbred less than 20 percent. Ralston e_t a1. (1948), studying the effect of crossing two inbred lines, observed the effect of in— breeding on production. They observed that outbred daughters of a sire produced 363 pounds of butterfat while his inbred daughters, inbred 12.5 percent, 25 percent, 31.25 percent, and 37.5 percent, produced 288, 269, 190, and 164 pounds of butterfat, respectively. The regression coefficient for an increase of each degree of in- breeding showed a decrease of 5.1 pounds of butterfat. Nelson and Lush (1950) found that production decreased as in— breeding increased. The regression of butterfat on inbreeding was -4.5 pounds per 1—percent inbreeding. It was possible in this herd to practice enough selection for production to counterbalance the effect of inbreeding and to raise the genetic ability for butterfat produc— tion approximately 40 pounds during the twelve years studied. Nelson suggests that if a breeding plan is followed in which the increase in intensity of inbreeding is less than 2 percent per gen—' eration, it should be possible to practice enough selection to coun- terbalance the decline in production expected to result from 198 inbreeding. Findings of Tyler e_t a1. (1946) bear out those of Nel— son. They found that milk production declined 73 pounds and but- terfat, 2.3 pounds, for each l—percent increase in inbreeding. They also found that the variation in the partial regression coefficients between sires was large enough that the offspring of some sires might be inbred as much as 25 percent without any apparent de— crease in milk and butterfat production. Swett e_t 11: (1949) found essentially the same results in their study of the results of the Beltsville inbreeding experiment. Production dropped approximately 17 percent in cows inbred more than 30 percent. Laben and Herman (1950) analyzed the effect of mild in— breeding by the intrasire regression of production on inbreeding. A significant decline of 66 pounds of milk and 2 pounds of butter— fat per 1—percent increase in inbreeding was observed. There was no significant effect on butterfat percentage. ‘ On the other hand, some investigations have shown that inbreed— i ing has no adverse effects on production. Woodward and Graves (1933), Bartlett and Margolin (1944), and Vainikainen and Nikkila (1946) have concluded that inbreeding does not adversely effect milk and butterfat production when superior animals are selected to start inbreeding with. Hays (1919), analyzing the inbreeding experiment m L" I- 198 inbreeding. Findings of Tyler e_t _a_1_. (1946) bear out those of Nel— son. They found that milk production declined 73 pounds and but—- terfat, 2.3 pounds, for each l—percent increase in inbreeding. They also found that the variation in the partial regression coefficients between sires was large enough that the offspring of some sires might be inbred as much as 25 percent without any apparent de— crease in milk and butterfat production. Swett gt 3.1. (1949) found essentially the same results in their study of the results of the Beltsville inbreeding experiment. Production dropped approximately 17 percent in cows inbred more than 30 percent. Laben and Herman (1950) analyzed the effect of mild in— breeding by the intrasire regression of production on inbreeding. A significant decline of 66 pounds of milk and 2 pounds of butter— fat per l—percent increase in inbreeding was observed. There was no significant effect on butterfat percentage. On the other hand, some investigations have shown that inbreed— ing has no adverse effects on production. Woodward and Graves (1933), Bartlett and Margolin (1944), and Vainikainen and Nikkila (1946) have concluded that inbreeding does not adversely effect milk and butterfat production when superior animals are selected to start inbreeding with. Hays (1919), analyzing the inbreeding experiment u f 199 onducted at the Delaware Experiment Station, found that a small mount of inbreeding—~under 10 percent—-—appeared to be associated [th the best production. Cows of the Rose May breeding——inbred the extent of 10 percent—~appeared to be heavier milkers and owed a higher percentage of butterfat than those of less inbreed— rth Weight In a progress report of inbreeding work carried on at the nois station to determine the possibilities in the development and ization of distinct superior lines of dairy cattle, Dickerson 10a) reported on the average effect of inbreeding on birth weight. res 16 percent inbred averaged nearly 10 percent lighter at birth did noninbred calves by the same sires after correction for ht differences due to sex and age of dam. This decline in L weight held for six of eight sires. Dickerson concluded that weight is determined to an important degree by the calves' -itance, since the dams of the calves in this report were not (1. Ljutikov (1944) analyzed the birth weights of 2,698 Bestuzev 8. One thousand, three hundred, and forty—two were inbred 12.5 percent as compared to 1,356 outbred calves. The inbred 200 eraged 7 percent lighter at birth than the outbred calves. Lgs are in close agreement with those of Dickerson (1940a). 1945), studying the consequence of close inbreeding at the Itate farm, ”Brody," found that the birth weight of inbred LS markedly less. - rler _e__t_ $1.: (1946) reported the extent and variability of in— effects in dairy cattle on birth weight by comparison of 1d outbred progeny of the same sires. They found that the regression of birth weight on 654 calves——adjusted for the of size of dam and sex——on their percentage of inbreeding ificant, with b = 0.28 pounds. Later work by Tyler e_t §_l_. l three unrelated herds of Holstein—Friesians bears out '1ier findings. The mean birth weight of 794 calves was ads and 85.5 pounds for outbred and inbred calves, re- y. Birth weight declined 0.28 pounds for each increase cent inbreeding, with sex of calf and size of dam held They concluded that the degree of heritability of birth rom additive gene effects in these data was about 50 to :nt. Voodward and. Graves (1946), reporting on a long—time in— ; experiment conducted by the Beltsville station, found that 201 of generations of inbreeding had a more pronounced ct on birth weight of calves than did the degree of in— .‘he weight of sixth—generation inbred calves was uni— than that of outbred calves, when the degree of inbreed— Lsidered. Nelson and Lush (1950) found that regression Lght on inbreeding of bulls and heifers showed a decrease th of a pound in birth weight for each increase of 1 inbreeding. On. the basis of this regression, a first- parent—offspring or full—sib mating would be expected verage birth weight by about 3 pounds. The effect of of the dam on birth weight of the calf was also studied. ssion of 'birth weight on inbreeding of the dam was —0.05 calves and -—0.14 for bull calves. The weighted mean of we a decrease of approximately one-eleventh of a pound for each l—percent increase in inbreeding. These find— similar to those of Ralston 31; 31. (1948), who found the n coefficient for an increase of one degree of inbreeding 14 pound. 1 the other hand, however, Bartlett e_t §_._l_. (1939), Bart- ._. (1942), and Margolin and Bartlett (1945), reporting on me inbreeding experiment by the New Jersey station, 202 that the degree of inbreeding or generation of inbreed— t adversely affect the birth weight of calves. Reports .in and Nikkila (1946) and Tyler _e_ta_l. (1949) are in harmony indings. Vainikainin and Nikkila, in studying the Ayrshire arge Finnish estate, descended from one bull and eight im- :rs, found 10 percent of the population to be inbred more cent. The birth weights of these inbred animals were no dif— n those of the outbred population. Tyler e_t _a._l., studying .s of unrelated cattle, designated all animals under 6 L outbred and those inbred over 6 percent as inbreds. 3 no significant difference between birth weights of inbred ad calves. Weight at Two Years and Over :lson and Lush (1950), studying the effects of mild inbreed— herd of Holstein—Friesian cattle, found the regression co— of weight on inbreeding increased up to two years of age. :hough the mean weight continued to increase, the regres— Eficients decreased with advancing age up to five years. icates that inbreeding not only resulted in lighter calves but slower gains during the first two years of life. After 203 :he inbreds apparently gained faster than the outbreds years of age, approached or exceeded the outbreds The outbred individuals merely approached their ma- 3 at a. younger age. These findings are Similar to those (1944), who found a definite depression in the develop— : from birth to eighteen months of age of inbred heifers :d to outbreds. er g al_. (1949) reported the effect of inbreeding on iairy cattle at eighteen months of age and at maturity. 1 that the regressions of weight on inbreeding at eighteen d at maturity were not significant. Swett g a,_l. (1949) lever, that at eighteen months of age inbred heifers were Ltely 12 percent lighter than outbreds and at maturity they rcent lighter than the outbreds, while Vainikainin and Nikkila uorted that inbreeding had no apparent effect on the conforma— 1rowth of the animals studied. Ralston e_t_a_l. (1948) found the 3 between the weight of inbred daughters—37.5 and over—— red daughters of the same bull to be significant at the t level. aker fig. (1945) found in the inbred daughters of one -Friesian sire that a significant proportionate decrease in 204 rred With an increased coefficient of inbreeding. These in harmony with those of Woodward and Graves (1946), :hat in the sixth generation of inbreeding the weight of two years of age was 20 percent below outbreds, and .ths after the fourth calving they were still 10 percent outbreds in weight. Bartlett e_t 3;. (1942, 1944) reported 1dings——that unless animals were more than 20 percent appreciable difference in weight of outbreds or inbreds 'ved. irtlett gt Q. (1942), at the New Jersey station, studied the of inbreeding on the type—rating of dairy cattle selected ee areas—«the New England States, New Jersey, and New They concluded that the inbreeding of a family of Holstein—- s produced animals that exhibited no relation between in— )f inbreeding and type classification. A total of 112 ani— LS classified in these findings. The later findings of and Margolin (1944) on further studies of the same family at, although animals with an inbreeding coefficient exceed- Opercent were inferior in size, their body type was superior 205 lose of the outbred controls. The report by Vainikainin and 11a (1946) is in harmony with the findings of the New Jersey 111. They studied the effects of inbreeding on the Ayrshire : population of a large estate in Finland. They concluded the type-ratings of the inbred individuals were as high as _of the outbred animals. Woodward and Graves (1946), at the Beltsville station, however, that, as inbreeding increased, the animals became 3h, acquired a rough hair coat, and were stubborn in appear— Their conclusions can be interpreted as a reduction in type ication of the inbred animals. Nelson and Lush (1950) re- the effects of mild inbreeding on type classification. They ed all animals each year, resulting in most animals being ed several times. The official terms of type classification ;ed, and it was intended that each class was divided into :bclasses: low, medium, and high-—thus resulting in a total zen type classifications. Heifers were also classified prior g. Each subclass was given a numerical value starting at low producers and going to 18 for high excellent. The score for the herd was 7, which is equivalent to a classi— f middle good. Data were available on the classification __-1 206 .5 individuals whose average inbreeding was 4.6 percent. They sired by thirty—seven different bulls, and the differences be— 1 groups of daughters from different bulls were not statistically 'icant. The intrasire correlation between inbreeding and type -0.12 and the regression of type on inbreeding was —0.04. the regression coefficient obtained here, it would require an Lse of inbreeding of 25 to 30 percent to lower the average lassification one—third of a class. 13: Research investigations and observations have indicated that y is affected by the genotype of mates. if a cow family ally inferior for good fertility were inbred, the homozygosity type of mating should increase the appearance of low fer- vhereas inbreeding could as well make a family homozygous :1 fertility. Several workers investigating the effects of in— g on dairy cattle have upheld these observations. Woodward and Graves (1946), in analyzing the inbreeding Holstein-Friesians at the Beltsville station, found that one tred 50 percent, and his daughters, inbred 37.5 percent, 1 a low conception rate, whereas the Conception rate of 207 r sires and their daughters was not affected. Akopjan :iting some cases of bad consequences of close inbreeding ussian state farm, "Brody," reported that the first conse- >f inbreeding was a loss of fertility. Matings of inbred :eded, on the average, not less than seven services in settle the cow, whereas outcrosses became pregnant, on age, with two matings. Vainikainin and Nikkila (1946) re— ~lat the inbreeding of Ayrshire cattle in Finland resulted in :rably lowered fertility rate. Bartlett gal. (1939) found that, , 'eral generations of inbreeding among Holstein—Friesians at Jersey station, the rate of fertility declined. Nielsen (1943), her hand, reported no decrease in the fertility rate when he ed Danish Milkrace cattle. Nielsen's findings are in agree— h the observations of Woodward and Graves (1946), who Lt certain inbred families had a fertility rate that was or superior to that of outbreds. ' of Calves 'oodward and Graves (1946) studied the mortality rate of lees as compared to outbred calves born between January to September 1, 1941. They observed 108 inbred female 208 and 41 outbred female calves born during the experiment. of the inbred calves died within a year, or 15 percent. ' the outbred calves died. They concluded that the inbred were not as vigorous as the outbred calves. Bartlett _e_t g. ound that the mortality rate among inbred calves was higher the outbreds. This increase in mortality rate was primarily the appearance of a lethal factor, ”Bulldog," and other body nations in one of the inbred families. Ljutikov (1944), stud— 142 inbred and 1,356 outbred matings found the mortality be 14 percent higher in the inbreds than in the outbreds. (1945) studied the effec'.s of inbreeding in Finnish Ayrshire nd found that the inland calves were less vigorous and showed .- mortality rate. nce of Lethal Defects n cattle, defects which cause death—-either prenatal or l——u.nder conditions of normal environment are referred to al defects." Other terms such as "sublethal," "semilethal," leterious characters" have been used by various reviewers ifying the observed mutants in dairy cattle. 209 The total loss due to inherited lethals is not known. They i constitute a portion of the calves born dead; this statistic )een reported by several workers. Linn _e_t a_1. (1949) have 'ted that approximately 5 to 6 percent of the calves born to is D. H. I. A. cows were born dead, in a study of 25,000 1gs. Miller and Gilmore (1949) reported a. death loss of 4, i 13 percent, respectively, for three Minnesota Agricultural 'iment Station herds. These authors attributed 4 to 5 percent loss as normal and 9 percent as due to a sex-linked lethal herd with the high death loss. In recent years a number of reviews have been presented ail which report the appearance of defects in domestic and animals. Among these are Hutt (1934), Eaton (1937), Lerner and Gilmore (1949, 1950, 1952). The most thorough and is that by Gilmore (1952), in which he provides numerous :s of the now known twenty-nine lethal defects in dairy Inve stigational P rocedure The data used in the present analysis of Red Danish cattle _._. and third—generation crosses) are the inbreeding coefficients, 210 ied by the application of Wright's (1922) formula, applied 3M procedure as outlined in a previous section of this The lifetime average of each cow was used as the mea— her producing ability. ’roduction records of milk and butterfat production and in— coefficients of 317 cows representing the progeny of ve sires were available. The production records of 122 d progeny of twenty—one of these sires were available for son purposes. ‘he amount of inbreeding in this population was due to the , coming from the same herd-—the parent Red Dane herd SDA. No special attempts were made to produce inbred The coefficients of inbreeding are nearly all due to the ing related and bred to related females by random chance. :fforts were made to move the bulls to the different bull n such order to keep inbreeding at a minimum. he correlation and regression analysis was conducted on ire basis with all calculations completed by an IBM pro— :xcept for the mechanics of constructing the analysis of :6 tables . 211 The data relative to the study of the appearance of lethal ects were determined from the cow—worksheet and calving— >ort forms. IBM Card No. 1 contained the calf data for each :shening of a cow as well as the sire and other details. These re summarized by sorting and tabulation. A six-generation pedigree was constructed for each reported thal defect in order to study the appearance of the possible trans— itters in each pedigree and to identify each sire as a trans— .itter. A breeding test was employed in which the expected or cal— ulated values were based on the assumption that 12.5 percent of he progeny should be defective when heterozygous sires were mated to progeny of heterozygous sires. The formula, [p + q]2 = [p2 + qu + qz], was used in order :0 determine the gene frequency of the defects in the American Red Danish cattle population and to test the randomness of mat— ings which resulted in the observations. 212 Investigational Res ults The Effects of Mild Inbreeding on Milk and Butterfat Production The amount of inbreeding in this population was low, since was due to the random mating of related females and bulls. 1though a total of 317 females were inbred, only 17 cows were Lbred over 10 percent. Of the cows with records, one had a co- fficient of 30 percent and four had coefficients of 25 percent. hese 317 inbred cows were sired by thirty—five sires. All of 1e inbred cows were either second— or third-generation Red ‘anish, since the first-generation females were the progeny of an utcross of two unrelated animals. The average inbreeding coef- cient for the inbred second- and third-generation females was 4 ercent and 7 percent, respectively. This shows that the relation—- hip existing between the sires used for this project increased, nd also that the inbreeding of the dams, second generation, con- ributed to the increase. Table XXIV shows the records made by cows sorted into ifferent inbreeding groups. From this grouping of production ‘ecords by inbreeding groups, it would appear that a. mild inbreed— ng of 1 to 5 percent caused an increase in average milk and -;—M 213 TABLE XXIV THE EFFECT OF INBREEDING ON PRODUCTION Inbreeding Avg. No. No. Avg. G Co f °f °f M’lk ”up 6 ' Cows Sires 1 0-0 -——— 122 21 8,690 0.01—0.05 0.03 213 31 9,213 0.06—0.10 0.07 87 28 9.157 0.11—0.18 0.13 12 6 9.924 0.25 + 0.26 5 4 11,047 TABLE XXIV (Continued) 214 Std. 6:? Avg. Std. Cgfef‘ Pct Dev Var Fat Dev. Var Fat 1,649 19.0 345 63 18.3 3.97 1 1,713 18.6 376 66 17.5 4.08 1,866 20.4 370 80 21.6 4.04 2,104 21.2 396 70 18.0 3.99 2,025 18.3 435 81 18.6 3.93 215 butterfat production. The differences for milk and butterfat are significant at the l-percent level of probability for this amount of data. There appears to be a slight decrease in milk and butterfat production with an increase of inbreeding coefficients for the group— ing of 6 to 10 percent. This is relatively small and is not signifi— cant at the l- or 5—percent levels of probability. The increases in the 11 to 13 and 25+ coefficients groups should probably be attrib— uted to the small number of observations in these groups, since an analysis of variance of the inbred daughters proved to be non— significant at either level of probability for butterfat production and barely significant at the 5—percent level for milk production. In the latter case, the statistic could occur, even by chance, one time out of twenty forla sample of this size. An intrasire correlation and regression of production on inbreeding anal ysis was conducted to determine the effects of in— breeding on production on an intrasire basis. These findings are presented in Table XXV. The regression estimates indicate that milk production decreased about 23 pounds and butterfat production decreased about 0.3 pound for each increase of 1 percent in in- breeding. The intrasire effect of inbreeding is shown more clearly in Table XXVI. The number of daughters in the various 216 TABLE XXV INTRASIRE CORRELATION AND REGRESSION OF PRODUCTION ON INBREEDING , Regression Record Correlation (lbs.) Milk Production —0.15 :1: 0.03 ~23.l Butterfat Production —0.05 i 0.02 -0.30 categories is rather small in some instances, since it was not de— signed during any stage of this project to obtain inbred and outbred daughters of each sire. The inbred and outbred progeny are all due to the random chance of bulls being bred to related or un— related females. It is evident from Table XXVI and substantiated with a highly significant "F" test by the analysis of variance that considerable variation exists between sires. Some sires' progeny performed better when inbred, whereas others showed a downward trend. From the intrasire regression coefficients shown in Table XXV, it appears that the over-all trend was downward; however, the variation between the progeny of the different sires was enough to keep this negative regression relatively small. TABLE XXVI 217 DAUGHTER'S PRODUCTION BY SIRES HAVING TWO OR MORE DAUGHTERS IN TWO OR MORE INBREEDING GROUPS Sire N°' °£ Inbreeding Avg. Avg’ Avg' No . Daugh— Group Milk Butte r— Pct. ters fat Fat 1557 7 .01—0.05 9,790 375 3.83 11 .06-0.10 8,246 326 3.95 2 .25 + 9,677 404 4.17 1558 18 .00 7,981 340 4.26 4 .01—0.05 8.152 361 4.42 5 .06—0.1o 8,355 350 4.18 1561 19 .01—0.05 9,581 376 3.92 5 .06—0.10 11,116 449 4.03 1564 3 .01—0.05 8,211 321 3.90 4 .05—0.10 9,863 423 4.28 1566 7 .00 8,882 362 4.07 6 .01—0.05 8,628 356 4.12 1568 3 .00 9,196 354 3.84 8 .01-0.05 8,753 367 4.19 8 .06-0.10 9,465 357 3.77 1569 3 .01—0.05 7,472 308 4.10 1.." —- .. .. 218 TABLE XXVI (Continued) Sire N°° of Inbreeding Avg. Avg‘ Avg' No. Daugh— Group Milk Butte r— Pct. ters fat Fat 8 0.06—0.10 8,164 328 4.01 1570 18 0.01—0.05 10,022 431 4.30 2 0.06—0.10 9,922 378 3.80 1571 6 0.00 8,918 360 4.03 16 0.01—0.05 9,354 363 3.88 1572 2 0.00 6,060 277 4.56 15 0.01—0.05 9,828 390 3.96 1573 3 0.00 8,279 338 4.08 24 0.01—0.05 9,548 375 3.92 4 0.06—0.10 9,065 373 4.11 1575 2 0.00 10,947 415 3.79 11 0.01~0.05 10,169 396 3.89 7 0.06—0.10 9,660 381 3.94 6 0.11—0.18 11,145 418 3.75 1578 11 0.00 8,865 350 3.94 9 0.01-0.05 9.528 388 4.07 2 0.06—0.10 9,688 375 TABLE XXVI (Continued) 219 Sire No. Of Inbreeding Avg. Avg. Avg. No. Daugh- Group Milk Butter— P ct. ters fat Fat 1579 3 0.01—0.05 8,382 316 3.76 3 0.06-0.10 8,216 350 4.25 1582 2 0.00 8,601 329 3.82 3 0.01—0.05 7,358 334 4.53 2 0.06—0.10 9.283 373 4.01 1607 12 0.01-0.05 9.179 393 4.28 3 0.06—0.10 9.665 431 4.45 I. 220 Lethal Defects The line—breeding and mild inbreeding practiced as a random occurrence in this project has brought to light two hereditary re— cessive defects in a number of the cooperating herds. One of the hereditary defects that has been segregated was a type of lameness or paralysis. This abnormality was present at birth and persisted until the death of the affected individual from other causes——usually at about two weeks. Inflammation of the kid— neys due to incomplete drainage of the bladder with urination and lack of Circulation in the extremities have been observed as the causative factors of death due to the animals' being unable to stand to allow proper drainage of the bladder and circulation in the legs. The affected calves lie on one side, usually with outstretched limbs and often with extremely bent hocks. The permanent lying Position leads to sores and loss of hair. None of the calves sur- Vived, although the owners improvised slings to allow the calves to stand” Even then, after six weeks or longer, no appreciable improVernent could be detected. Attending veterinarians were the first to suspect that the con dition might possibly be an inherited anomaly, since all therapeutic atte 11'1th were of no avail. Several of the calves affected with this ¥ #- 220 Lethal Defects The line—breeding and mild inbreeding practiced as a random occurrence in this project has brought to light two hereditary re— cessive defects in a number of the cooperating herds. One ef the hereditary defects that has been segregated was a type 0f lameness or paralysis. This abnormality was present at birth and persisted until the death of the affected individual from other causes——usually at about two weeks. Inflammation of the kid— neys due to incomplete drainage of the bladder with urination and lack 0f Circulation in the extremities have been observed as the causative faCtO rs of death due to the animals' being unable to stand to allow Proper drainage of the bladder and circulation in the legs. The a1'fected calves lie on one side, usually with outstretched limbs and often with extremely bent hocks. The permanent lying position leads to sores and loss of hair. None of the calves sur— vived, although the owners improvised slings to allow the calves t ° Stand- Even then, after six weeks or longer, no appreciable im 1- p ovement could be detected- A “eh-ding veterinarians were the first to suSpect that the Condition might possibly be an inherited anomaly, since all therapeutic atte t mp 5 We re of no avail. Several of the calves affected with this 221 condition were brought to the outpatient clinic at the Veterinary Center of Michigan State College; no response to treatment was observed by the clinicians. X rays were also made which showed no gross deformities of the vertebrae or limbs (Langham, 1952), The author succeeded in obtaining two affected calves from widely separated herds—~290 miles apart-—and observed them for several weeks and found the symptoms to be the same as reported by the breeders in whose herds these anomalies appeared. Gross and histological studies were made of the spinal cord and brain tissues; the results are not available at this writing. Figures 7 and 8 show photographs of these calves and the gross appearance of the anomaly. Sixty—five afflicted animals——thirty—eight heifers and thirty- seven bulls-—have been reported by twenty-seven dairymen in all sections of the state. Five—generation pedigrees were constructed of each defective calf, and it was found that these sixty—five calves were sired by seventeen Red Danish sires. From a study of these pedigrees, it was found that four additional sires could be classi— fied as heterozygous for the paralyzed condition, since they appeared as the sire of a first—generation dam which produced the defective calf. In order to determine the kind of inheritance pattern these 222 Figure 7 Plate A illustrates the voluntary movements of the forelimbs and body with incoordinate voluntary movements of the rear limbs. Plate B shows the calf in a supported upright position with the forebody supported by the forelegs. The hind limbs, however, demonstrated no ability to support the weight of the body. A typ— ical deviation of the tail is illustrated which snaps back into place after being straightened passively. 224 Figure 8 Plate A illustrates the typical lying position which this calf assumed without being able to move other than raising its head, in— coordinately moving the fore and hind limbs, and jerkily moving the tail. Plate B illustrates the inability of the calf to sustain its own weight, especially with the hind limbs. This was more pro— nounced in this calf than the one in Figure 7. 226 data fitted, a summary was madeef the progeny of these sires when mated to daughters of heterozygous sires. This is presented in Table XXVII. The expected values are based on the assumption that 12.5 percent of the progeny should be defective. This was an excellent agreement between the observed re— sults and the theoretical expectations. Judging from the results of this study, this lethal condition -—paralyzed hindquarters——is conditioned by a single autosomal recessive gene. The second hereditary defect observed during the course of this investigation was one of a type of mummification or anky— losis of the neck and limbs. The affected calves were born at full term or two weeks premature. They exhibited a shortened, thickened, and stiffened neck and stiffened legs, sometimes bent backwards with the joints ankylosed and distinctly thickened. The hoofs and skin were well developed. In many cases veterinary aid was necessary, since in all cases parturition was extremely difficult and often resulted in permanent sterility of the dam due to injury of the uterus. Attending veterinarians called the dairy—- men's attention to a similar condition reported in Danish cattle in Denmark . 227 TABLE XXVII THE OBSERVED AND EXPECTED RESULTS OF MATING SIRES HETEROZYGOUS FOR THE PARALYZED CONDITION TO THEIR OWN DAUGHTERS AND TO DAUGHTERS OF OTHER HETEROZYGOUS SIRES Normal Defective Calves Calves Percent Total Observed 566 65 12.24 531 Expected 464.63 66.37 12.50 Difference +1.37 —l.37 —00.26 2 . x = 0.0004. Forty—two of these ankylosed fetuses occurred in the herds of eleven dairymen. In the majority of the cases, the sex of the defective fetuses was not observed. However, in the instances noted, it appeared evenly divided between male and female. Four- teen sires were found to have sired these defective fetuses, thereby being heterozygous for the defect. From the construction of pedi— grees of these affected fetuses, it was found that four additional sires could be classified as heterozygous for the defect since they Were sires of first—generation females producing offspring with the anomaly. #— 228 A summary of the progeny of these sires mated to their own daughters and the daughters of the other heterozygous sires is presented in Table XXVIII. The expected or calculated values were based on the assumption that 12.5 percent of the progeny should be defective. Table XXVIII shows that the difference be— tween the observed and expected frequencies was exceedingly small. From these findings it can be concluded that a single auto— somal recessive gene conditions the lethal defect—-mummification. In the analysis for the gene frequencies of the two defects, an analysis of the paralyzed condition is shown. Let p 2 frequency of Pr (normal) and q = frequency of pr (paralyzed). Then, p + q = 1 (P+<1)2=P2+2PQ+<12=1 q2 = 0.022 q = V0.022 = 0.149 p = 1 _ 0.149 = 0.851 p2 = 0.7242 zpq = 0.2538 2 p + qu + q2 = 0.7242 + 0.2538 + 0.022 1.00 229 TAB LE XX VIII THE OBSERVED AND EXPECTED RESULTS OF MATING SIRES HETEROZYGOUS FOR THE ANKYLOSED CONDITION TO THEIR OWN DAUGHTERS AND TO DAUGHTERS OF OTHER HETEROZYGOUS SIRES 12:13:? sz:1:t::e Pe rcent T otal Observed 301 42 12.24 343 Expected 300.13 42.87 12.50 Difference +0.87 —0.87 ——00.26 2 x = 0.0005. The above analysis suggests that 25 percent of the American Red Danish cattle are heterozygous for the paralyzed condition. An analysis of the observed occurrence (1.4 percent) of the mummified defect suggests that 11 percent of the population are heterozygous for the defect. The Chi—square test was employed to determine if matings had been random or whether the occurrence of the defects was due to selection of mates. Since x2 = 0.004; therefore, it can be concluded that the matings were random. 230 Other Effects Calf Mortality No attempt was made to determine the number of calves that died within one month of birth, which would have been a partial measure of the vigor of the inbred calves. However, the number of calves dead at birth was recorded in order to ascer— tain the possibilities of one of several lethals of this type being found in this population. Table XXIX shows the percentage calf mortality of the various foundation breeds and of the crossbred females. It is interesting to note that the percentage calf mor— tality increased with each crossbred generation. This increase could not be attributed to the occurrence of a lethal defect other than the mummified condition as described earlier in this paper. The evidence of calf mortality did not increase as the degree of inbreeding increased. Further study is needed to ascertain if a factor or factors are responsible for the increase of calf mor— tality with successive generations of Red Danish sires. TABLE XXIX 231 CALF MORTALITY AT BIRTH FOR THE FOUNDATION BREEDS AND CROSSBRED GENERATIONS Breed Percent Percent Calves Total Calves Inbreeding Dead at Birth Observed Guernsey 0.00 2.33 425 Holstein 0.00 1.42 410 Shorthorn 0.00 0.69 291 Jersey 0.00 1.37 136 Mixed 0.00 1.78 225 First cross 0.00 3.71 1,700 Second cross 0.04 7.73 913 Third cross 0.07 9.34 333 m 232 Sterility Three categories of sterility were included among a number of other reasons for disposal of a cow. Percentage figures were determined for the percentage of the sterility of total disposals. The three categories of sterility were: (1) bred four times or more in normal heat periods, but did not settle; (2) complete lack of heat periods; and (3) apparently settled, showed signs of heat three or more months later. Table XXX shows the percentage of cows by combined foun— dation breeds and by crosses that were disposed of for one of the three types of sterility. The rate of incidence of sterility was no higher in inbred cows than in outbred animals. Further study is required to determine if this increased sterility in successive crosses is due to an inherited condition. Discussion and Conclusion The Effects of Mild Inbreeding on Milk and Butterfat Production The regression of production on inbreeding observed in this study is something less than that found by several investigators——Nelson and Lush (1950), Laben and Herman (1950), Tyler gt fl- (1949), and 233 TABLE XXX PERCENT OF COWS OF THE FOUNDATION BREEDS AND CROSSES DISPOSED OF BECAUSE OF STERILITY Percent Percent Disposal Total Inbreeding Due to Sterility Observations Foundation breeds 0.00 13.1 469 First cross 0.00 20.23 420 Second cross 0.04 25.8 256 Third cross 0.07 28.8 65 Ralston gt a_l_. (1948). On the other hand, the present observations are in close conformity to those of Hays (1919), Woodward and Graves (1933), Bartlett and Margolin (1944), and Vainikainin and Nikkil'a (1946), where production performances of inbreds——not over 10 percent inbred—-were superior to those of outbreds. One should expect the results of these observations to dif— fer somewhat from those of other reports, since the genetic com— position of the cattle observed in this study is certainly of a dif- ferent nature. The reported literature dealing with the regression of production on inbreeding indicated that the animals observed in those studies were either purebred or high—grade animals of many 234 generations of breeding of a single line and strain. The inbred animals in this observation carry either one—fourth or one—eighth blood of an entirely different breed than the ancestry responsible for the coefficients of inbreeding. It is rather surprising that even greater differences did not occur. The intrasire effect of inbreeding on milk production, tab- ulated in Table III, illustrated that some sires can stand the tests of inbreeding, while others cannot. It is evident that considerable variation between sires as to the effects of inbreeding exists. Tyler e_t a}; (1949) found this to be true in their study. There also appeared to be some differences in response by progeny of the different inbreeding groups within a sire. However, since these records are expressed on a single population basis——many herds together-«some of the differences between the inbreeding groups between and within sire progeny groups must be attributed to dif- ferences in the expression of the genotypes of the animals due to the variation in environmental conditions existing between herds. The intrasire regression of inbreeding on production have been adapted to correct for the variations between animals due to genetic differences between sires by most of the workers reporting on the effects of inbreeding on production, and also with the assumption ‘y a?!" 235 that most of the daughters of a sire are in one herd and exposed to the same environmental effects at the same time. Today, how-— ever, as in this study, it is not unnatural to find daughters of a sire in many herds and expressing their genetic composition at Widely separated intervals due to the use of sires in artificial breeding studs. The writer recognizes that these findings would be more nearly accurate if analyzed on an intraherd, intrasire basis; however, it was beyond the scope of this study to obtain those statistics. Future studies of this aspect should be directed in that direction. Lethal Defects The appearance of two lethal defects became apparent after a preliminary investigation of reports by dairy farmers cooperating in the Red Danish project; namely, that numerous calves were born too weak to stand alone, and others had enlarged, stiffened joints and were dead at birth. These same conditions have been reported in the Red Dane Milkrace cattle in Denmark by Loje (as cited by Hutt, 1934). His observations were of limited number and did not entirely fit the one—factor hypothesis. Later reports on the para— lyzed condition in Red Dane Milkrace cattle were made by Nielsen 235 that most of the daughters of a sire are in one herd and exposed to the same environmental effects at the same time. Today, how— ever, as in this study, it is not unnatural to find daughters of a sire in many herds and expressing their genetic composition at widely separated intervals due to the use of sires in artificial breeding studs. The writer recognizes that these findings would be more nearly accurate if analyzed on an intraherd, intrasire basis; however, it was beyond the scope of this study to obtain those statistics. Future studies of this aspect should be directed in that direction. Lethal Defects The appearance of two lethal defects became apparent after a preliminary investigation of reports by dairy farmers cooperating in the Red Danish project; namely, that numerous calves were born too weak to stand alone, and others had enlarged, stiffened joints and were dead at birth. These same conditions have been reported in the Red Dane Milkrace cattle in Denmark by Loje (as cited by Hutt, 1934). His observations were of limited number and did not entirely fit the one-factor hypothesis. Later reports on the para— lyzed condition in Red Dane Milkrace cattle were made by Nielsen A— ‘9 1.1 236 (1942) and Rasmussen (1943). Neilsen reported that the first ob— servation of this condition in Red Dane cattle was made in 1924. Up to 1941, the total observations made were 294, sired by thirty- nine bulls with the frequency (reported in 1941) being fifty—six paralyzed calves. Nielsen found that his data did not fit the one- factor hypothesis clearly—-that it appeared that two factors condi- tioned the anomaly. However, Rasmussen (1943) found that the . data did fit the one—factor hypothesis, which is the same as that found in the present study. A similar condition, observed by Drimmelen (1942) and Tuff (1948) in African and Norwegian Red Poll cattle, respectively, have been reported. Tuff's findings of ninety normal and fifteen lame calves in 105 observations of heterozygous bulls mated to daughters of heterozygous bulls substantiated the present conclu- sions that this paralyzed condition is caused by a single recessive autosomal gene. The ankylosed condition found in the Danish cattle by Loje (as cited by Hutt, 1934) has not been reported recently. However, Tuff (1948) found a very similar condition in Norwegian Red Poll cattle. His observations were rather limited, but his conclusions were that the condition was caused by a single recessive autosomal 237 character. Gilmore (1952) made mention of similar observations that have been made in Guernsey cattle in the United States. How— ever, to the writer's knowledge, similar conditions of either the paralyzed or ankylosed anomalies have not been reported for other breeds of cattle not reviewed here. To dispel the possibility that these conditions were caused by environmental factors, attention should be called to the appear— ance of these conditions in numerous herds located in different sections of the state. It would seem impossible that only herds of Red Dane cattle in those sections would be deficient in some essential vitamin or mineral. Then, too, many of the paralyzed calves have had therapeutic treatments by attending veterinarians in the field or Veterinary Clinic of Michigan State College. The e"idence that these two factors——paralyzed hindquarters and mum— rnifi"lation, as originally named by Loje (as cited by Hutt, 1934) ——are conditioned by a single recessive autosomal character is the Cl°seness of fit of the expected to the observed results in Tables XXVII and XXVIII. That these two conditions were not carried by the founda— tion breeds used in this project is evidenced by the fact that of 3’854 reported calvings from Red Danish sires crossed with the g 238 foundation breeds, no defects were observed. The first observa— tion of either defect was the result of matings of a Red Danish sire to a first-cross female. Othe r Effe cts Calf mortality and sterility increased with successive gen— erations of purebred Red Danish sires. The incidence was not correlated with the degree of inbreeding. Further study is needed to establish if inherited factors are operating. These findings do not agree with those previously reviewed. Summary A study was made on the effect of inbreeding in the Amer- ican Red Danish cattle population. The effect of inbreeding on milk and butterfat production, appearance of lethal characters, calf mor— tality, and sterility was investigated. The second—cross animals Were the first to have inbreeding coefficients, since the first-cross Were the progeny of an outcross. The inbreeding of the inbred ani— mals of the second cross averaged 4 percent, while those of the third cross averaged 7 percent. 239 The analysis of the effect of inbreeding showed that the inbred animals increased milk and butterfat production 3 and 8 percent, respectively, over the noninbred animals. This indicated that the percent of butterfat increased with inbreeding which the analysis revealed as true. However, within the inbred groups it was found that for each l—percent increase in inbreeding, milk and butterfat production decreased 23 and 0.3 pounds, respectively. This study revealed the occurrence of two lethal defects—~— PaI'a-lyzed hindquarters and mummified fetuses. The number of observations of both defects fits the expected Mendelian ratio of 12-5 percent occurrence when heterozygous sires were mated to daughters of heterozygous sires. Therefore, it was concluded that bOth lethal characters are inherited independently as single auto— somal recessive factors. Additional study is advisable to determine if the observed inc rease in calf mortality and sterility with successive generations Of PUI’ebred Danish sires is an inherited condition. THE INFLUENCE OF THREE NONHEREDITARY FACTORS: TIME TREND AND YEARLY ENVIRONMENTAL CHANGES. MONTH OF CALVING, AND CALVING INTERVAL ON MILK AND BUTTERFAT PRODUCTION Review of Lite rature Introduction It is important to know the extent to which differences be— tween production records are caused by environmental factors. The more pronounced ones are age at calving, times milked daily, and length of lactation. Many investigators, however, have investi— gated the effect of other environmental factors such as year—to— year variations, time trends, calving interval, season of freshen— ing, and length of service period on milk and butterfat production. The results obtained in these studies have been varied, although it appears that the most important of the minor environmental fac— tors are year-to—year variations and time trends, length of dry period, and length of calving interval. The present study was undertaken to measure the relative importance of year—to-year variations and time trends, length of 241 dry period, and length of calving interval as causative factors of variation among records of the American Red Danish cattle pOp- ulation. Year—to-Year Variation and Time Trend Plum (1935) studied Iowa Cow Testing Association records from ninety—five herds of Guernsey, Holstein, and Jersey cattle. He found that during the eleven years included in that study, the average production increased about 50 pounds of butterfat. How— ever, only 2.8 percent of the total variance is due to yearly changes. The influence of year—to—year changes within a herd accounted for 6 percent of the total variance or roughly 10 per- cent of the intraherd variance. These findings were somewhat lower than those found by Patow (1930), who seemed to consider year-to-year variation as the major environmental cause of vari- afion. Laben and Herman (1950) analyzed the effect of yearly variation and time on the total variance of production for the Missouri Station Holstein—Friesian herd from its foundation in 1902 to January 1, 1950. They found a significant upward time trend in production. Differences between five-year periods accounted for L 242 5.5 percent of the total variance in milk production, 20.4 percent of the total variance in butterfat production, and 38.8 percent of the total variance in butterfat percentage. The portion of vari- ance attributed to butterfat and butterfat percentage was due to continued effort to select high—testing sires with essentially the same milk production as the herd average. Chai (1951) analyzed the production records of three Hol— stein herds owned by Michigan State institutions. He found the portion of intraherd variance due to difference between years to be 4.9 percent, while that portion due to total variance was 3.7 percent. He utilized Hoel's simplified runs method for analyzing time trends and found no significance. Influence of Month of Calving on Milk and Butterfat Production and Butterfat Percentage There are several factors contributing to the apparent vari- ations in production caused by different seasons of freshening. Numerous observations have been made on the effects of these factors on production characteristics. It has been shown by Turner (1927a) that a rapid increase in milk and butterfat production oc— curs immediately following freshening, and reaches a peak level during the first one and one-half to two months, and then declines 243 to the time of cessation. The height and maintenance of the peak determines the total production. Therefore, any environmental factors affecting the persistency of the peak increases the slope of the declining lactation. If a cow freshens at an optimum month, so that the peak can reach and maintain its maximum height and persistence, and then toward the end of her lactation receives another favorable environmental boost, she will produce more than if freshened at any other time of the year. From management and economic consideration there may be reason for having cows freshen in different seasons; particu— larly either fall or spring. However, the main objectives of the present study have been to test statistically whether or not there is any significant relation between the month of calving and the production in the lactation period following it. Effect of Season of Calving Ragsdale and Brody (1922a) analyzed three kinds of records—- Guernsey data on a national scale, Jersey data from a single state (Missouri), and Holstein data from a single herd. They concluded that best production records were obtained when cows calved in September or October. The percentage of fat in milk when plotted 244 followed a general curve—-being lowest during the summer months, then gradually rising, reaching a peak during the winter months, and then declining during the spring and summer. Ragsdale and Brody (1922b), observing ten cows over 41 days (March 13 to April 22) at the Missouri station, noted that with mean daily temperature ranges between 27° and 700 F., there was a depression of almost 0.2 percent in the average fat content of the milk for each increase of 100 F. between the observed temperatures. Hays (1926), studying the entire Missouri station dairy herd under uncontrolled conditions for 258 days, from January, 1924, ob- served a 0.079-percent fat decline with each 100 F. rise in mean temperature. Temperature, Hays concluded, is a major factor in the seasonal variation in percentage of fat in cow's milk. Eckles and Shaw (1913) concluded that the effect of season is apparently the result of weather conditions, especially heat and humidity. They found that during a period of hot, humid weather the per— Centage of fat was depressed, while during dry and cool conditions the test is increased. Baker and Cranfield (1933), studying the average fat percent at collecting depots in England, observed a bMOW—normal fat percent for the summer months of 1925 to 1929 inclusive, as compared to a twenty—year average. They found this 245 decrease to be correlated with a below—normal rainfall for those years. This agrees with the findings of Turner (1923), who found that the principal variations in milk production in regard to sea— son of calving were the changes in pasture during the Spring and summer months. However, Wylie (1925), in a study of more than 2.900 R. O. M. records of Jersey cows, found that cows with the highest milk production did not necessarily show a seasonal effect and that July was an optimum month. Cannon (1933) analyzed 68,000 Iowa Cow Testing Associa— tion records to observe season of calving effect on subsequent pro—— duction.. He grouped the records into months of calving and found that cows freshening in November gave the highest yields. In the November to June groups, the ones freshening each month had a larger production than the group of the preceding month. Cannon developed a set of correction factors to be used when comparing cows' records starting in different months. He suggested that these can also be applicable to the North Central states. Headley (1933), studying the records of Nevada Agriculture Experiment Station herd, observed from a total of 124 records that cows freshening before March averaged 40 pounds daily with a gradual decline for those starting in later months until the average for 246 September was 32 pounds. Following September, a gradual in— crease was found until a peak of 40 pounds was reached in De— cember. Plum (1935) assembled records of ninety-five herds of the Iowa Cow Testing Association between 1922 and 1932, comprising 6,900 records of three breeds—«Holstein, Jersey, and Guernsey, both purebred and grade. He analyzed these data to determine the causes of the variations in cows' records and observed that cows calving from November to January produced 13.6 percent more butterfat than cows calving from May to July. Although the season of the year has a significant influence upon production, it is a minor cause of total variation—~measuring only 3 percent of the total. Woodward (1945) tabulated 15,422 records of cows in dairy—herd improvement associations from twelve states. His findings are similar to numerous other investigators. He found July the lowest yielding month, with November the highest, but that the difference was not very great. He concluded that it was evident that the total quantity of milk produced in a lactation period is not a factor of major importance for the dairyman to consider in determining the best time to have his cows freshen-— if he is able to provide adequate feed for his herd at all seasons L_- _ 247 of the year. Frick and co—workers (1947) analyzed 22,212 Con- necticut Dairy Herd Improvement Association lactation records and found very similar data to that of Plum (1935). They found Feb- ruary the most favorable month of freshening, which exceeded by 13.7 percent the least favorable month, July. This association of milk yield with month of calving was similar for each of four breeds studied: Guernsey, Holstein, Ayrshire, and Jersey breeds. On the other hand, two groups of workers, Oloufa and Jones (1948) and Arnold and Becker (1935), working with Oregon and Florida data, respectively, found no difference in total pro- duction due to season of calving. Both groups of workers con- cluded that the difference in seasons was not of enough variance to cause seasonal differences. Dickerson (1940) analyzed Wiscon— sin Dairy Herd Improvement Association records in an attempt to determine the relative genetic worth of corrected and cow records. He calculated each cow's records in terms of five kinds of records: 240-day, 305-day, 365—day, total lactation, and testing year. He concluded that season of calving had little effect on records of about 305 days' duration, since only 4 percent of the total variance could be attributed to season of calving with only 1.9 percent for records of about 305 days' duration. 247 of the year. Frick and co—workers (1947) analyzed 22,212 Con— necticut Dairy Herd Improvement Association lactation records and found very similar data to that of Plum (1935). They found Feb— ruary the most favorable month of freshening, which exceeded by 13.7 percent the least favorable month, July. This association of milk yield with month of calving was similar for each of four breeds studied: Guernsey, Holstein, Ayrshire, and Jersey breeds. On the other hand, two groups of workers, Oloufa and Jones (1948) and Arnold and Becker (1935), working with Oregon and Florida data, respectively, found no difference in total pro— duction due to season of calving. Both groups of workers con— cluded that the difference in seasons was not of enough variance to cause seasonal differences. Dickerson (1940) analyzed Wiscon— sin Dairy Herd Improvement AssociatiOn records in an attempt to determine the relative genetic worth of corrected and cow records. He calculated each cow's records in terms of five kinds of records: 240—day, 305—day, 365—day, total lactation, and testing year. He concluded that season of calving had little effect on records of about 305 days' duration, since only 4 percent of the total variance could be attributed to Season of calving with Only 1.9 percent for rec0rds of about 305 days' duration. 248 Chandrashaker (1951) analyzed the production of butterfat in relation to the month of calving in five breeds-—Holstein—Friesian, Jersey, Brown Swiss, Ayrshire, and Guernsey——in the Michigan State College dairy herd. By using the analysis of variance method of analysis, he found the "F" test between months to be nonsignifi— cant. He concluded that the month of freshening had no appreciable and significant influence on the yearly butterfat production. Chai (1951) analyzed the butterfat production records of three Michigan State institutional herds composed of Holstein—Friesian cattle and found the effect of month of calving to be significant in two of the three herds. He found the highest butterfat production for the month of March, it being significant from any other month. About 2. percent of the total variance was accounted for by the different months of calving. Effect of Low Temperature That the dairy cow withstands long periods of exposure to temperatures as low as zero degrees Fahrenheit with little loss in production has been shown by nmnerous workers. Buckley (1913) at Maryland found that there is no greater variation in the flow of milk from cows exposed to a temperature 249 of —14 degrees in an open stable than in those cows in a closed stable which were exposed to a temperature of +11 degrees. He concluded that the effect of low temperatures is practically nega— tive in reducing the flow of milk. Davis (1914) observed similar results in an "open—shed housing versus closed—stable" experi— ment under Pennsylvania conditions. He also concluded that the drOps in atmospheric temperature did not affect the cows in the open shed any more than it did those in the closed stable. Kelley and Rupel (1937), studying the relationship of en— viromnental conditions to milk production under Wisconsin con— ditions, found that 500 F. appeared to be about the optimum tem- perature for dairy cows; however, temperatures below this op— timum did not decrease production after the first three milkings following the fluctuation. Dice (1940) found that cows turned out into a snow-filled yard during the day maintained as good a pro- duction as those kept in. Cows that Were bedded down in an open shed maintained production as well as those cows housed in a closed stable. He concluded from his observations that milk cows can stand exposure to cold temperature and that they pro- duce as well in an open stable as they will when temperature is maintained at 500 F. An explanation published by Armsby (1917) 250 eXplains why the cow can resist extreme low temperatures without detrimental effects to production. Using experiments by Kellner, he Showed that of the energy evolved as heat by a fattening steer on full feed, about 62 percent of it was used to maintain body heat and 38 percent was surplus at optimum temperatures. To bear -this out, Armsby cited work by Jordan (1908) to show that the heat PrOduced by one dairy cow was greater by 85 percent and for an— other was 54 percent greater than that amount needed for main- tenance and that "so far as mere maintenance of body temperature goes, no reason appears why a cow might not be subjected to com— Paratively low temperatures without causing any increased ketabolism for Sake of heat production solely." Ragsdale e_t E1. (1950) found, hoWever, that milk yield begins to decline earlier in Jersey cows at temperatures below 400 F. than it did in Holstein cows. The critical low temperature for Jerseys seemed to be 80 F., whereas at this temperature the Holstein showed little effect-~if any-— in decline of milk yield. E ~%f High Tempe rature . o Temperatures in excess of 85 F. have, however, a marked d . etrlmental effect on milk production as found by Rhoad (1935) in 250 explains why the cow can resist extreme low temperatures without detrimental effects to production. Using experiments by Kellner, he showed that of the energy evolved as heat by a fattening steer on full feed, abOut 62 percent of it was used to maintain body heat and 38 percent was surplus at optimum temperatures. To bear -this out, Armsby cited work by Jordan (1908) to show that the heat produced by one dairy cow was greater by 85 percent and for an— other was 54 percent greater than that amount needed for main- tenance and that "so far as mere maintenance of body temperature goes, no reason appears why a cow might not be subjected to com- paratively low temperatures without causing any increased ketabolism for sake of heat production solely." Ragsdale _e_t 31. (1950) found, however, that milk yield begins to decline earlier in Jersey cows at temperatures below 400 F. than it did in Holstein cows. The critical low temperature for JerSeys seemed to be 80 F., whereas at this temperature the Holstein showed little effect——if any—— in decline of milk yield. Effect of High Temperature . o Temperatures in excess of 85 F. have, however, a marked detrimental effect on milk production as found by Rhoad (1935) in 251 Brazil. He observed that nine Holsteins imported into the tr0pics, produced on a Cornell ration only 56 percent of the corrected average production of their dams and two granddams as given on the pedigree sheets. That this reduction in apparent capacity was due to continuous high temperature is borne out by observations of Villegas (1939). He observed that Holstein cows imported into Singapore and kept in an air-conditioned barn at 700 F. averaged 24 pounds of milk a day as compared to 9 pounds a day for a similar group in an open, well—ventilated barn, exposed to tropical temperature 5. Regan and Richardson (1939) have shown, under controlled conditions, that as atmospheric temperature increased from 400 to 95° F., milk production gradually dr0pped from 29 to 17 pounds a day. With water buffaloes, Sinha and Minett (1947) have shown that cooling of the animals by splashing with cool water gave a significant increase in yield of milk over a nonsplashed group. On the other hand, Miller gt 2.1. (1951) reported that there appeared to be no relationship between the amount of milk produced and the Sprinkling of Jersey and Holstein cows, although the sprinkling was effective in cooling the cows as evidenced by decreasing respira- tion rates and body temperatures. However, the authors suggested | :- 252 that further studies should be conducted with the temperature con—— siderably above 800 F. This bears out some of the observations of Riek and Lee (1948), who reported that Jersey cows subjected to temperatures ranging from 850 to 1100 F. twice weekly for 7 hours demonstrated no effect in milk yield and butterfat percent— age. Weaver and Matthews (1928) observed that, as temperature increased both inside and outside of the barn, fat percentage de— clined. They found that the amount of variation in fat percentage declined. They found that the amount of variation in fat percent- age due to outside temperature changes ranged, according to a regression equation, from 0.0017 for the Ayrshire to 0.0103 for the Guernsey, and that for inside temperature changes ranged from 0.0023 for Holsteins to 0.0148 for Jerseys. Ragsdale and Brody (1922a, 1922b) found that the percentage of fat in milk, when plotted, follows a general curve, being lowest during the hot summer months, gradually rising and reaching a peak during the winter months, and then declining during the spring and summer. Ragsdale _e_t Ei' (1948, 1949, 1950, 1951), in a series of ex— periments at the Missouri laboratory, found that as temperature increased over 500 F., milk production declined. These workers \ 253 0 concluded that 80 to 850 F. appeared to be the critical tempera- ture for milk production. That there are breed and Species differences in response to elevated temperatures have been shown by Hammond (1931) in observations made in the countries of Jamaica and Trinidad. He Observed that the Shorthorn and Ayrshire breeds were the least able to withstand the high temperatures as measured by milk pro- duction. It was also noted that the Jersey, Holstein—Friesian, and Red Poll breeds appeared more adaptable——in that order. These observations noted by Hammond were borne out by the findings of Rhoad (1 936), who has shown that when European and Indian cattle were Crossed, the resulting crosses yielded more milk than their E “rcpean dams under Brazilian conditions. Regan and Richardson 1 3 1 9 8): Comparing the heat tolerance of the Holstein, Jersey, and Cu emsey breeds, found that as the mean temperature was increased by 10° F . - ' ' ' . increments, there was a small decline in daily milk prod - “Ct-Ion. Depending on breed, a sharp decline was observed when 80° 0 . . . to 85 F. was reached, With the Holstein cows show1ng IeSs heat tolerance than the Jersey and Guernsey cows. Seath and Min er (1947) compared the heat tolerance of Jersey and Holstein Q0ws by using correlations on an intraherd basis and found that m 254 the Holstein cows had a significantly lower heat tolerance than did the Jersey cows. Gaalaas (1947), observing seventy—four Jersey cows for five years, which involved 137 heat-tolerance coefficients, made the fol- lowing observation. Not much change in the average heat tolerance of the herd occurred from year to year; a definite difference in heat tolerance existed in the different age groups; heat tolerance coefficient is a reasonable stable individual characteristic at the age of four years and above, but not at two or three years of age; there is a real difference in physiological response of different cows to the same environmental conditions as measured by body temperature. He also found that the stage of lactations and gesta~ tion has little, if any, effect on the heat tolerance, and that there is some difference in the reSponse of groups of daughters of various sires when measured at two and three years of age, but little, if any, when measured at four years or more of age. Ragsdale _e_t El. (1948, 1949, 1950, 1951) observed a breed and species difference in the critical high temperature with ref— erence to milk production. They observed that 750 to 800 F. was the critical high temperature at which the depressing effect on milk production became evident for Holstein cattle, 800 to 850 F. 255 for Jersey and Brown Swiss cattle, and 900 to 950 F. for Brahman cattle. Influence of Calving Interval on Milk and Butterfat Production The variations in production records due to this portion of environment can be attributed to two factors working within this classification—~that part of the variation due to length of service period after freshening, and that part due to length of dry period preceding the lactation. Numerous workers have found that the latter months of ges— tation have an inhibitory effect upon the production of milk. Ga— vin (1913a) found that if cows were bred five to eight weeks after calving,~a sharp drop in milk production occurred from the twenty-— first to twenty—fourth week after calving. Whereas cows bred from seventeen to twenty weeks after calving did not show the cha -1.cter— istic drop until the twenty-ninth to thirty-second week after :alving. Eckles (1923) reported a 19—percent decrease in annual milk pro— duction for a group of cows bred to calve within twelve months as compared to the same group when bred to calve at more than fif— teen months' interval. 256 Hammond and Sanders (1923) found that a short service period reduced total milk yield whereas a long service period increased it. These observations imply that a cow early in gesta— tion during a lactation will produce less than one that is late in gestation during a lactation period. Hammond and Sanders sug— gested correction factors to apply to records of cows either af- fected or not affected by gestation. They found that in cows con— ceiving within nineteen days after calving, a correction of 30 per— cent should be added to the total milk yield. These correction figures decreased with each nineteen-day interval of increased service period up to ninety to one hundred days after calving, which is 0 percent correction. For cows not serviced until after one hundred days after calving, they suggested a minus cor— rection factor to be applied for each nineteen-day interval up to 8. -11 percent for 160- to l79—day service period. These obser— vations were borne out by Sanders (1923). He devised a factor for lactations which he called the "Shape Figure" and found a positive correlation between the "Shape Figure" and length of the service period. As the length of the service period increased, the ”Shape Figure" increased. Sanders also observed a positive 257 correlation between height of the ”Shape Figure" and the total lactation yield. Brody _e_t_ _a_l_. (192.3), studying the records of Advanced Reg— istry Guernsey cows, found that cows kept farrow until within five months of the end of the lactation produced more than cows bred so that more than five months of gestation appeared in the lacta— tion period. Gaines and Davidson (1926), studying 4,552 records of the Advanced Register of the American Guernsey Cattle Club, found results very similar to those of Brody e_t El. (1923). Gaines and Palfrey (1931), using 352 Red Danish cows with ten uninter— rupted calvings, found the average calving interval to be 401 days. They found a positive correlation between the calving interval and total yield. Hammond and Sanders (1923) found that the length of dry period had a great effect on subsequent production to a certain point. Any length beyond this changed total production slightly. They observed that a dry period length from 0 to 39 days caused a l3—percent reduction in total yield; 40 to 79 days dry period caused a 2.5—percent reduction in yield; 80 to 119 days were con- sidered as having no effect on subsequent yield, and a dry period of 120 days plus caused only a Z-percent increase in total yield. 258 Dickerson (1940), studying Wisconsin D. H. I. A. cow records, ob- served that 6 to 9 percent of the total variance was to the preced— ing dry period. However, he concluded that preceding dry periods had little effect on 305—day records. Klein and Woodward (1943) studied.15,442 lactation records of D. H. I. A. cows from twelve States. They found that cows dry one to two months gave 9.2 per— cent more milk than when dry zero to one month; cows dry two to three months gave 4.3 percent more than when dry one to two months and 13-5 percent more than when dry zero to one month. Cows dry three to four months gave 14.9 percent more than cows dry zero to one month, They concluded that a dry period of fifty—five days was the optimum length for cows calving at twelve-month intervals. Chai (1951) studied the effects of calving interval on the same lactation and the next lactation. He found the effect of dif— ferenCe lengths of calving intervals to be quite large. Over 14 Pei-Cent of the total variance was credited to different lengths Of Calving intervals for the same lactation and 3 percent of the total Variance due to its effect on the following lactation. He concluded that the total variance due to different calving intervals totaled 18 percent of environmental effects. He developed a set of cor— recu ' on factors baSed on twenty—day intervals which appeared to 259 be as important as correcting for age at freshening and even more so than corrections for length of lactation. Investigational P rocedure The survey forms which have been used to collect data for the investigation of the genetics of the Red Danish cattle included appropriate Spaces to include the year and month of freshening and calving interval relative to each production record reported. One hundred and eighty-eight dairy farmers, c00perating in the Red Dane project and testing monthly under supervision of the Dairy Herd Improvement Association, cooperated in completing the Survey forms needed for this study. Eighteen counties were rePreSerited in the survey located in the northern, eastern, south- em, and western section of the state. All production records were corrected to a 2X, 305—day, M’ E' basis. A cow freshening in any particular month was credited to that month. The same was true for year of produc— tion. It was felt that since a considerable number of records was invol Ved, the systematic errors would cancel each other. For the year Of Production study, three general groupings were made of the b reeds involved. The Holstein—Friesian, Jersey, and Brown L ,. 260 Swiss breeds were pooled into Group I; the Mixed, Shorthorn, and Guernsey breeds into Group II; and the first—generation, second- generation, and third-generation graded—up Red Danish crosses were placed into Group III. The purpose of this grouping was an attempt to remove some of the variance due to differences be— tWe n herd management. It was beyond the scope of this study t0 analyze nine breeds for 188 herds on an intraherd basis. Since the ”t' ' test had shown the differences between the mean produc~ tion Of the breeds within each group to be nonsignificant, it was felt that some of the similarity between the mean production of the breeds, as grouped, was due to contemporary herd environ- ments. The calving interval was calculated by thirty—day class intervals. If a calving interval extended beyond the class mark It was placed into the next class. With enough numbers involved as In this study, the systematic errors should cancel each other. Since all of the data had been punched on IBM cards, the procedure followed to extract the desired data for this study was one of sOrting and tabulation involving card number 3. The method used in this study for analyzing the effect of Yearl y environmental changes, calving interval, and month of 261 freshening was the unequal subclass numbers analysis of variance given by Snedecor (1950). For determining time trends in data there are several methods of analysis available. The method used here, after Hoel (1948). and recently used by Chai (1951) on production records, is the so~called "runs"-——a kind of test of randomness of sequences. The median average for each group was determined as outlined by Snedecor (1950), and the weighted mean yearly average for each group Was also determined. The same was completed for the groups When c3C>Ihbined. The average production for each year for each group and for the three groups together were assigned the letter "a" if they were less than the median, and the letter "b" if they were greater than the median for their respective groups. It was felt that for the yearly environmental changes and time trend analysis that the mean butterfat production would show the necessary results and it would be unnecessary to test the mean milk production for these factors. 261 freshening was the unequal subclass numbers analysis of variance given by Snedecor (1950). For determining time trends in data there are several methods of analysis available. The method used here, after Hoel (1948), and recently used by Chai (1951) on production records, is the so—called "runs"-——a kind of test of randomness of sequences. The median average for each group was determined as outlined by Snedecor (1950), and the weighted mean yearly average for each group was also determined. The same was completed for the groups When combined. The average production for each year for each group and for the three groups together were assigned the letter "a" if they were less than the median, and the letter "b” if they Were greater than the median for their respective groups. It was felt that for the yearly environmental changes and time trend analysis that the mean butterfat production would show the neCessary results and it would be unnecessary to test the mean milk production for these factors. 262 Investigational Results Time Trend and Yearly Environmental Changes The average yearly butterfat production figures for each of the three groups are given in Table XXXI. Figure 9 shows these figures in graphic form. Tables XXXII and XXXIII show that there was a highly sig- nificant difference between different years for butterfat production either for a single group or after combining the three groups. There are indications that the averages of the last three years may have a time trend since they are considerably higher than the averages of the earlier years. However, the averages for the last year in each group is somewhat lower than the preceding two—- indicating that the high averages for the years 1949 and 1950 may have been due to yearly variations. The portion of the intragroup variance which was accounted for by differences between years was Slightly above 3 percent, while the portion of total variance due to differences between years was close to 7 percent. In analyzing the time trend of these- data the following four Sets of arrangements were computated: m L a TABLE XXXI 263 A LIST OF YEAR AVERAGES OF BUTTERFAT PRODUCTION Group One Group Two Group Three Total Year No. No. No. No. of Avgs. of Avgs. of Avgs. of Wtd. Rec— Rec— Rec— Rec— Avgs. ords ords ords ords 1939 17 335 11 335 ——— ——— 28 335 1940 26 352 21 328 ——— ——— 47 341 1941 34 390 28 329 ——— ——— 62 362 1942 20 385 17 300 19 363 56 352 1943 28 354 40 328 54 334 122 337 1944 26 376 51 333 71 350 148 349 1245 22 348 53 312 118 363 193 347 1946 32 351 43 312 157 364 232 353 1947 51 373 71 313 216 342 338 341 1948 82 390 137 328 325 371 544 363 1949 71 412 171 349 452 379 694 375 195° 53 421 132 354 451 390 636 385 1951 29 400 58 347 250 375 337 372 —\ Weighted Aver-fige 384 335 372 364 —\ 264 Figure 9 Average yearly butterfat production of the Holstein—Friesian; Jersey, and Brown Swiss breeds (Group I); the Mixed, Shorthorn, and Guernsey breeds (Group II); and the first-, second—, and third— generation Red Danish cross females. t t54 .3? .SE 3k 2:. O _ _ _ _ _ A . _ _ _ _ ._ O m “M \\/// O Q - x . 8 9 2n \ // n u man 8% V 1 m 7 deem. she M w d a U M o M. 2m 8.5 M 0 3 N u. o M 8m 88 N v 4... m. on. _ . _ . _ I _ . _ 274 Figure 11 Frequency of Calvings for the Different Months I L. 2 o t 0.3 xok k we Haw“. 63‘ >43» 3.35 x <2 .«1< £54 fink <<<5 . - u _ _ _ _ _ _ _ § enN 8 M own HINON 83d SSNIA'IVJ JO UJQNIHV .8, 276 then they rise rather rapidly from August to October, thence slowly to December, and then drop slowly to January. Table XXXV shows that the month effect on milk and butter- fat production was significant (P < 1%). The portion of total vari— ance due to month effect on milk and butterfat production was 1.9 percent and 2.4 percent, respectively. According to the level of difference needed for significance between averages of milk production, the means for the months of October, November, December, January, February, and March were not significantly different. They did differ significantly from the other months. The means for April and September were not sig— nificantly different, however. They differed significantly from the winter months and from the summer months of May, June, July, and August. The differences between means of the summer months were not significant. Therefore, from analyzing the means with the "t“ test (P < 0.05), it was concluded that the means for milk production were grouped into a peak of six late fall, Winter, and early spring months, with the low points for production being in the summer months and a month of transition lying between the high and low points. 277 TABLE XXXV ANALYSIS OF VARIANCE OF YEAR EFFECT ON BUTTERFAT PRODUCTION OF EACH GROUP De— Pro— Source grees Sums of Mean duction 0f Of S uar s S uares F Variance Free— q e q dom Milk Total 3,446 18,985,052,670 5,509,300 Between 11 416,908,000 37,900,727 7.01** months w‘thm 3,435 18,568,144,670 5,405,470 months Butterfat Total 3,446 23,358,895 6,779 Between 11 631,494 54,409 8.67** months W‘ . ”hm 3,435 22,727,401 6,616 months ** (P < 0.01) POI'tion of total variance due to month effect: Milk: 5,509,300 — 5,404,570 _ 5,509,300 ' 1'97" Butte rfat: 6.779 _ 6,616 \——— = 2.4% 6,779 278 The means of butterfat production showed a slightly differ— ent trend. A peak occurred for October, November, and December, but the difference between means was not significant (P < 0.05). The differences between the means of these months and the other months were significant. The low mean months, May, June, July, and August, did not exhibit significant differences between their means. As in the case of milk production, there were transition months on either side of the high~ and low—mean months. From the tests of significance, the peaks in butterfat production were very similar to those of milk production. This was expected, since total milk production and total butterfat production were sig- nificantly correlated. The frequency of calvings for the different months are shown both as numbers per month and percentage of total observed. The largest frequency of calvings occurred during the month of March, with the frequencies for the months of October, November, Decem— ber, January, February, and April somewhat similar but below that of March. The lowest frequency occurred in June, followed closely by July and August. These same data are shown graphically in Figure 11. 279 Effect of Calving Interval on Milk and Butterfat Production Table XXXVI shows the distribution of the records and average production for each calving interval. These same figures for milk and butterfat production are shown graphically in Figure 12. These data show that the effect of the calving interval on the next lactation was rather pronounced. As the calving interval lengthened, there was a steady in— crease in average milk production. This increase was linear until the interval 440 to 469 was reached. Thereafter for the next in— tervals plotted, the milk production leveled off and seemed not to be increased or decreased by a longer interval. The same general trend for butterfat production is exhibited by Table XXXVI and Figure 12. However, the mean butterfat pro— duction as affected by the calving interval reached its maximum height one interval later——47O to 499—“than milk production. In the next two intervals it declined. From Table XXXVI and Figure 12, it was concluded that the calving interval for securing best mean production for the next lac- tation was a calving interval of 440 to 469 for milk production and 470 to 499 for butterfat production. TABLE XXXVI 280 DISTRIBUTION OF RECORDS AND THE MEAN MILK AND BUTTERFAT PRODUCTION OF NEXT LACTATION FOR THE DIFFERENT CALVING INTERVALS Records Production f Inlt):r;;al % of Mean 0;, of Mean Z7181; No. Total Milk Max. 1 Butter— Yield2 Yield fat 290—319 139 7.59 8.457 86.49 348 87.43 320-349 423 23.09 8,699 88.97 358 89.94 350—379 573 31.28 8,933 91.36 371 93.21 380~409 333 18.18 9,230 94.49 380 95.47 410-439 152 8.30 9,491 97.07 387 97.23 440-469 85 4.63 9,816 100.03 394 98.99 470-499 47 2.56 9,763 99.85 405 100.50 500-529 50 2.73 9,817 100.04 399 100.03 530—559 '30 1.64 9,712 99.33 390 97.98 \‘m 1 An average of the last four intervals. 2 An average of the last three intervals. 281 Figure 12 The Effect of Length of Calving Interval on Milk and Butterfat Production of the Subsequent Lactation 8.3.3 3.42:3 62435 36.2.... 5.68 :15. are: .26? 43.6% .362. 366% 24.6% o_1 _ _ _ _ _ _ _ _O m m \1 ommv ml ohm 8.8 3% 62.8 ('59?) NOIJJHOOUd 117.78.71.1/79 8.8.8 one e83 ('597) NOIJ. anooad mm 283 An analysis of variance——Table XXXVII—~was calculated to find Whether the effect of calving interval on milk and butterfat production for the next lactation was statistically significant. The results shown in Table XXXVII were highly significant. The por— tion of the total variance due to differences caused by calving in- terval effect was 3.0 percent and 3.2 percent for milk and butter—- fat production, respectively. By use of the "t" test——to test for significance of the dif— ( ferences between the mean milk and butterfat production for the ‘ various intervals, it was found that the differences were not sig— nificant until two intervals or sixty days had elapsed. This was true for both milk and butterfat production. However, beyond the interval of 440 to 469 the differences between the means were not Significant. Table XXXVIII shows the average calving intervals for the various breeds and crosses in this study. As can be seen, the differerl(:es between the mean intervals for the various breeds and cr ' 08338 are not significant. 284 TABLE XXXVII ANALYSIS OF VARIANCE OF EFFECT OF CALVING INTERVAL ON MILK AND BUTTERFAT PRODUCTION OF THE NEXT LACTATION De— Pro— Source rees duction of g of Sum of Mean F Char- Variance Free— Squares Squares acter dom Milk Total 1,832 9,025,973,505 4,926,841 Between , 8 301,649,490 37,706,186 7.88** intervals ,W‘thm 1,824 8,724,324,015 4,783,072 intervals Butterfat Total 1,832 13,958,253 7,619 Petween 8 502,571 62,821 8.51** intervals ,W‘thm 1,824 13,455,682 7,377 intervals ** (P < 0.01) Milk: 4,926,841 — 4,783,072 _ 4,926,841 ‘ 3'07“ Butterfat: 7,619 - 7,377 7,619 4 3‘27" 285 TABLE XXXVIII AVERAGE CALVING INTERVALS SHOWN BY MONTHS AND CLASS INTERVALS FOR THE VARIOUS BREEDS AND CROSSES STUDIED Breeds Class Interval Month Avg. Number Guernsey 350—379 12.2 350 Holstein 380-409 12.7 384 Shorthorn 380—409 12.5 341 Jersey 380—409 12.8 167 Brown Swiss 380-409 12.6 48 Mixed 380-409 13.1 187 First cross 380—409 12.6 1,207 Second cross 380—409 13.2 557 Third cross 380—409 12.7 98 Average 380—409 12.7 Total 3,339 286 Dis cus sion and Conclusions Time Trend and Yearly Environmental Changes The factors which account for this effect such as climate, economical changes, feed conditions, all have a direct or indirect effect on milk and butterfat production of dairy cows. These fac— tors are the ones that primarily cause year—to—year fluctuations in production averages. The effects of these are usually cancelled out, if the data extends over a reasonable period of time. How- ever, factors such as better feeding practices, better housing, better management practices, all have a direct or indirect effect that cause what is known as a time—trend effect. When analyzing data, the time—trend effect can bias the analysis much more than the year—to-year fluctuations. In these data there was a time trend toward higher production during the later years which was only slightly significant. Even though a slight time trend existed in these data, the animals composing each group which were compared for various traits were distributed over the entire period considered, thus essentially eliminating any bias. The portion of the intragroups variance which was accounted for by difference between years was 3 percent. The portion of the 287 total variance which was accounted for by difference between years 'was 6.7 percent. These figures are in close agreement with Chai (1951) and Dickerson (1940). Effect of Month of Calving on Milk and Butterfat Production ' The results of the effect of month of calving on milk and butterfat production in this study coincide roughly with findings of other workers who observed a seasonal effect. As would be ex-- pected, the months showing maximum effects were different because of different geographical regions differing in temperatures, rainfall, pasture conditions and other environmental factors. It is felt that the findings of this study differed from the re-— sults of the two investigations of this nature in Michigan due to sampling and environment. The studies of Chandrashaker (1951) and Chai (1951) were with data from one and three herds, respec— tively, while these data for the present study were from 186 herds in all sections of the state. Chandrashaker (1951) studied the effect of month of calving on butterfat production of the Michigan State College herd and found the portion of variance due to month effect nonsignificant. That his findings are reasonable is accepted, since the college herd had an undoubtedly standardized environment and 287 :al variance which was accounted for by difference between years 3 6.7 percent. These figures are in close agreement with Chai '51) and Dickerson (1940). Effect of Month of Calving on Milk and Butterfat Production ‘ The results of the effect of month of calving on milk and :erfat production in this study coincide roughly with findings of :r workers who observed a seasonal effect. As would be ex— :ed, the months showing maximum effects were different because lifferent geographical regions differing in temperatures, rainfall, ure conditions and other environmental factors. It is felt that the findings of this study differed from the re— : of the two investigations of this nature in Michigan due to 51mg and environment. The studies of Chandrashaker (1951) Chai (1951) were with data from one and three herds, respec— y, while these data for the present study were from 186 herds 1 sections of the state. Chandrashaker (1951) studied the effect onth of calving on butterfat production of the Michigan State ge herd and found the portion of variance due to month effect gnificant. That his findings are reasonable is accepted, since allege herd had an undoubtedly standardized environment and 288 ort is made to maintain it as such. Chai (1951), in in— .g the environmental and genetic factors affecting Holstein— cattle in three state institutional herds found that month ; had a significant effect on mean butterfat production in a three herds studied. The high month for mean production :h, with June and August, the low months. The peculiarity aults was that the mean production for the month of July ficantly higher than those for June and August. He at— his to either sampling errors or to certain management that compensated for the usual adverse conditions for ing in that month. His findings of the mean production 1 as being the peak production month is in disagreement indings of the present study in which December appears ak month. nnon (1933) analyzed 68,000 Iowa Cow Testing Association ) observe month—of—calving effect on subsequent production. November as the peak month with June, July, and August I months. Plum (1935), studying similar data for ninety— , combined 6,900 records of three breeds in Iowa and ted the earlier findings of Canon. Woodward (1945) 15,000 D. H. I. A. records from twelve states into 289 s of calving and found July to be the month of lowest yield on subsequent lactations and November the highest. From y of Table XXXIV and Figure 10, it can be readily seen that dings of the present study were in almost complete agree—- rith the findings of Cannon, Plum, and Woodward. There was :eption in these data, however. The peak month was De— Although the maximum production for both milk and but— vas reached in December, the means are not significantly t from those of November. Essentially then, these data agreement with those of other workers studying D. H. I. A. T. A. records. The portion of total variance due to month effect found in ly compares very favorably with the findings of other work-— .ai (1951) attributed 2 percent of the total variance be ob— 1 his data to month effect. Dickerson (1940b), in analyz- onsin D. H. I. A. records, found that 4 percent of the Lance observed could be attributed to month of freshening, 1.9 percent when all records were of 305—day duration, present study, while Plum (1935) found that 3 percent of variance was caused by month—of—freshening effect. The if variance of the data in the present study, Table XXXV, 290 t the portion of total variance due to month—of—freshening 5 1.9 percent and 2.4 percent for milk and butterfat pro- respectively. nother aspect (of minor importance to the present study, terest from a. general point of View) which grew out of the E—freshening study was the frequency of calvings per month. Ld of information has not been previously ascertained for section of dairy cattle in Michigan. Chandrashaker (1951) d on his findings in the Michigan State College herd and .0 difference in frequency of calvings for the different months year. Chai (1951) found significant differences in the fre- of calvings for the three state institutional herds. He con- that the frequency was not controlled to the maximum 3 for production. The frequencies during the low months were ;h or higher than the peak production months. In the present . it was found that the frequency of calving followed th«- trend .lk production. The calving frequency for the highest month 'oduction was nearly twice that of the low month of production. mgh it must be added that the month with the highest calving uency in these data~—March-——was not the highest production 11:; it was a secondary peak month. 291 he reasons that calving frequencies are not more closely ed with maximum months of freshening for production in .ta could be due to various reasons: (1) farmers' ignorance ffect of month on production and consequent lack of attempts ‘ol the time of freshening, (2) milk markets such that farm— st supply a certain amount of milk during specific months :ive a much lower price for extra milk during flush periods, roductive troubles may cause cows not to settle at the proper nd (4) dairying is a year—around business from the invest— .f capital concerned and it is of the best interest for a ran to be milking cows twelve months of the year. Enough nce in production means for several of the months existed vhen comparing small numbers of animals such as in bull , it may be advisable to take into account the month of ning for the various lactations, if those data were available 3 population under study. In the present study the widest :nce between means was in the amount of 12. percent, which make quite a difference in a dam and daughter comparison. animal calved in the low month and the other in the high , the month effect alone would make a difference of 48 pounds on a 400-pound level. 292 Effect of Length of Calving Interval on Milk and Butterfat Production The data observed in this study compare favorably with :f Dickerson (1940). He observed that 6 to 9 percent of 3.1 variance could be attributed to the preceding dry period. records were calculated to 305 days in length, the variance reduce approximately 50 percent. A long calving interval 3 in an increased production in the following lactation s, 1923; Gaines and Palfrey, 1931). In this study, 3 and rcent of the total variance of milk and butterfat production, :tively, was attributed to the length of preceding calving in— Hammond and Sanders (1923) found that a dry period length :0 39 days caused a l3—percent reduction in total yield. For iterval of 290 to 319 in this study (Table XXXVI) which cor— nds roughly to 0 to 39 day interval of Hammond's and San- study, it was found that total yield was reduced by 13 and :rcent for butterfat and milk production, respectively, as and to the calving interval of 440 to 469 days which gave mum production. Chai (1951) proposed correcting the butterfat 293 :ion by 14 percent when the preceding calving interval was 319 days. When comparing the production records of a relatively few s—-—such as five dam—daughter pairs, it would be well to take :count the length of preceding dry periods. When large num— re involved, it would probably become unnecessary since the ad short intervals would tend to balance. Summary A slight time trend was evident in these data. Although the {es of all categories of cows reached a high peak the last :ars, there was a sharp drop in the last yearly average. The analysis of variance showed that 7 percent of the total Lee could be attributed to differences between years. Since :tion records were obtained from foundation breeds during the periods as those of the crosses, it was concluded that the r changes and time trend, did not bias the data; consequently, was no necessity of a correction for time trend. The analysis of the effect_of month of calving showed that freshening during the months of May, June, July, and August .ced from 10 to 12 percent less milk and butterfat. The cows 294 renting in the month of December produced 1,055 and 47 :13 more milk and butterfat, respectively, than did cows fresh— ; in June. An analysis of variance of these data attributed oximately 2 percent of the total variance of milk and butterfat uction to month—of—freshening effect. However, when comparing .1 numbers such as dam-daughter comparisons in a bull proof, suggested that the month of freshening be taken into account. An analysis of the effect of calving interval on the subse— t lactation showed a significant increase in milk and butterfat uction as the interval lengthened. With intervals of thirty , it was found that milk and fat production increased 3 to 4 ent with each increase of one interval. Cows with 290 to 319 calving intervals produced approximately 13 percent less milk butterfat than did cows with 440 to 469 day calving intervals. analysis of variance attributed 3 and 3.2 percent of the total ance of milk and fat production, respectively, in these data to Length of calving interval. Q‘ .‘l LITERATURE CITED W. W. 5 Heredity and Milking Functions; Jour. Dept. Agr. Victoria (Australia), 24:1—8. .n, K. A. 5 Posledstvijs teanago inbridingn. Sovhog. Proigvad., 8/9: 44—46. (Consequences of Close Inbreeding.) Anim. Breed— ing Abstr., 13—14leO. N. N., , L4 A Standard For Evaluation of Dairy Sires Proved in Dairy Herd Improvement Associations. Jour. Dairy Sci., 27: 835—847. ny, Ernest L. 51 Unpublished Manuscript. my, Ernest L. 52 Personal Communications. sby, H. P. 117 The Nutrition of Farm Animals. The Macmillan Co., New York. P. 454. dd, P. T., and R. B. Becker. 935 The Effect of Season of the Year and Advancing Lactation Upon Milk Yield of Jersey Cows. Jour. Dairy Sci., 18: 621—27. (RepOrt from Florida.) ell, S. A. 945 Breeding Systems Used to Produce the Highest Yielding Guernsey and Holstein Cows. Jour. Anim. Sci., 4:146—50. .er. A. G., and H. T. Cranfield. [933 Variations in Composition of Milk in Certain Midland Dis- tricts of England During the Years 1923-31. Jour. Dairy Res., 4:246—54. I‘ll 296 er, G. A., S. W. Mead, and W. M. Regan. .945 Effect of Inbreeding on the Growth Curves of Height at Withers, Weight and Heart Girth of Holstein Females. Jour. Dairy Sci., 282607—10. er, Marvel L., L. N. Hazel, and C. F. Reinmiller. 943 The Relative Importance of Heredity and Environment in the Growth Rate of Pigs at Different Ages. Jour. Anim. Sci” 2:3—13. er, T. A. 926 Transmission of Butterfat Percentage by Holstein—Friesian Sires. Delaware Agr. Exp. Sta. Bul., 145:1-5. zer, A. C. 952 Personal Communications. Llett, J. W., and S. Margolin. 944 A Comparison of Inbreeding and Outbreeding in Holstein— Friesian Cattle. New Jersey Agr. Exp. Sta. Bul., 712.. lett, J. W., R. P. Reece, and O. L. Lepard. 942 The Influence on Birth Weight, Rate of Growth and Type of Dairy Cattle. Jour. Anim. Sci., l(3):206-—12. lett, J. W., R. P. Reece, and J. P. Mixner. ’39 Inbreeding and Outbreeding Holstein—Friesian Cattle in an Attempt to Establish Genetic Factors For High Milk Pro— duction and High Fat Tests. New Jersey Agr. Exp. Sta. Bul.,.667. er, R. B., and P. T. Dix Arnold. *35 Influence of Season and Advancing Lactation on Butterfat Content of Jersey Milk. Jour. Dairy Sci., 18:389—99. V, J. C. 45 Reliability of Averages of Different Numbers of Lactation Records for Comparison Dairy Cows. Jour. Dairy Sci., 28:355—66. 297 :y, J. C., and Jay L. Lush. 939 High Records Contrasted With UnSelected Records and With Average Records as a Basis for Selecting Cows. Jour. Dairy Sci., 222607—17. iy, Samuel, A. C. Ragsdale, and Charles W. Turner. 923 The Effect of Gestation on the Rate of Decline of Milk Secretion With the Advance of the Period of Lactation. Jour. Gen. Phy. 52441-777. uet, Paul, and Jay L. Lush. 947 Heritability of Amount of Spotting in Holstein—Friesian Cattle. Jour. Heredity, 38:99—106. :ley, S. S. 913 Open Stables Versus Closed Stables For Dairy Animals. Maryland Agr. Exp. Sta. Bul. 177. .on, C. Y. )33 Seasonal Effect on Yield of Dairy Cows. Jour. Dairy Sci., 16:11—15. Le, W. E. 119 Inheritance of Quantity and Quality of Milk Production in Dairy Cattle. Proc. Nat. Acad. Sci., 52428—34. Chen Kang. ’51 The Relative Importance of Genetic and Environmental Factors on the Butterfat Production of Holstein-Friesian Cattle. Unpublished Ph.D. Thesis. Michigan State Col- lege Library, East Lansing, Michigan. drashaker, Beereppa. 51 Genetic and Enviromnental Contributions to the Economic Characteristics of the Michigan State College Dairy Herd. Unpublished Thesis, Ph.D. Michigan State College Library, East Lansing, Michigan. nittee 0n Investigations. 40 Amer. Soc. Anim. Prod. Proc., p. 378. tau 298 land, Lynn. '31 The Contribution of the Dam in Inheritance of Milk and Butterfat. J. Dairy Sci., 14:379—93. land, Lynn. '38 The Use of Records in Evaluating the Inheritance of Cows and In Proving Bulls. Jour. Dairy Sci. 21:651—60. ton, F. E., and Dudley J. Cowden. '49 Applied General Statistics. Prentice—Hall, New York. ison, F. A. '25 Measuring the Breeding Value of Dairy Sires by the Rec— ords of Their First Few Advanced Registry Daughters. Illinois Agr. Exp. Sta. Bul., 270. s, H. P. '14 The Effect of Open—shed Housing as Compared With the Closed Stable for Milch Cows. Pennsylvania State Col- lege Ann'l Rep., 183—226. J. R. '40 The Influence of Stable Temperature on the Production and Feed Requirements of Dairy Cows. Jour. Dairy Sci., 23:61. :rson, G. E. 403. Effects of Inbreeding in Dairy Cattle. (Progress Rep.) Jour. Dairy Sci. 23:546-47. :rson, G. E. 40b Estimates of Producing Ability in Dairy Cattle. Jour. Agr. Res., 61:561-86. :rson, G. E. 41 Estimates of Producing Ability in Dairy Cattle. Jour. Agr. Res. 61:561-86. :y, H. C., and Pedro Labarthe. 45 Predicting the Transmitting Ability of Young Dairy Sires For Milk Production, Butterfat Test and Butterfat Pro— duction. Jour. Dairy Sci., 28:893-900. 299 Drimmelen, G. B. Van. 1942 Lameness in a Stud Cattle Herd Possibly of Hereditary Origin. Jour. S. Afr. Vet. Med. Assoc., 13:105—10. Anim. Breeding Abstr. 11:159. Eaton, O. N. 1937 A Summary of Lethal Characters in Animals and Man. Jour. Heredity 282320-26. Eckles, C. H. 1923 Dairy Cattle and Milk Production. Macmillan Co., New York. Eckles, C. H., and R. H. Shaw. 1913 The Influence of the Stage of Lactation on the Composi- tion and Properties of Milk. U. S. Dept. Agr. Bur. An. Indus. Bul., 155:1—88. Edwards, Joseph. 1932 The Progeny Test as a Method of Evaluating the Dairy Sire. Jour. Agr. Sci. (England) 222811—37. Eldridge, F. E., and G. W. Salisbury. 1949 The Relation of Pedigree Promise to Performance of Proved Holstein-Friesian Bulls. Jour. Dairy Sci., 32: 841—48. Fohrman, M. H. 1943—47 Crossbreeding Dairy Cows. Year Book of Agr., 177—84. Fohrman, M. H. 1947 Progress Report on Crossbreeding of Dairy Cattle at Beltsville. Jour. Dairy Sci., 30:551. Fohrman, M. H. 1949 U. S. Dept. Agr. B. D. F. M. Inf. 30 (Supplement). Frick, George E., A. I. Manns, and Steward Johnson. 1947 The Relation of Seasons of Freshening to Milk Produc- tion. Jour. Dairy Sci., 30:621-40. 300 Gaalaas, R. F. 1947 A Study of Heat Tolerance in Jersey Cows. Jonr Dairy Sci., 30(2):79—85. Gaines, W. L. 1936 Correction Factors and Germ Plasm in Dairy Cattle Breeding. Amer. Soc. Am. Prod. Proc., 50—52. Gaines, W. L., and F. A. Davidson. 1926 Rate of Milk Secretion as Affected by Advance in Lacta— tion and Gestation. Univ. of Illinois Bul. No. 272. Gaines, W. L., and J. R. Palfrey. 1931 Length of Calving Interval and Average Milk Yield. Jour. Dairy Sci., 14:294—306. Garkavi, O. U. 19.;5 Problems in Breeding Work With Grade Cattle. Anim. Breeding Abstr., 15:172. Gavin, William. 1913a Studies in Milk Records: The Influence of Foetal Growth on Yield. J. Agr. Sci., 51309—19. Gavin, William. 1913b Studies in Milk Records: On the Accuracy of Estimating a Cow's Milking Capability by Her First Lactation Yield. J. Agr. Sci., 52375—90. Gifford, Warren. 1930a The Mode of Inheritance of Yearly Butterfat Production—- An Analysis of Holstein—Friesian Sires. Missouri Agr. Exp. Sta. Res. Bul., 144. Gifford, Warren. 1930b The Value of a Sire. Guernsey Breeders' Jour., 38:259. Gifford, W., and C. W. Turner. 1928 The Mode of Inheritance of Yearly Butterfat Production. An Analysis of the Progeny Performance of Ayrshire Sires and Dams. Missouri Agr. Exp. Sta. Res. Bul., 120, 1-52. Gilmore, Lester O. 1949 The Inheritance of Functional Causes of Reproductive In- efficiency: A Review. Jour. Dairy Sci. 32:71-91. Gilmore, Lester O. 1950 Inherited Non—lethal Anatomical Characters in Cattle: A Review. Jour. Dairy Sci. 33:147—65. Gilmore, Lester Odell. 1952 Dairy Cattle Breeding. F. B. Lippincott Co., New York. P. 595. Goodale, H. D. 1927 A Sire's Breeding Index With Special Reference to Milk Production. Amer. Nat. P. 641. Goulden, C. H. 1949 Methods of Statistical Analysis. John Wiley 8: Sons, Inc. Chapman and Hall, London. Gowen, John W. 1920a Inheritance in CrOSSes of Dairy and Beef Breeds of Cattle. Transmission of Milk Yield to the First Genera- tion. Jour. Heredity 112300-16. Gowen, John W. 1920b Inheritance in Crosses of Dairy and Beef Breeds of Cattle. III. Transmission of Butterfat Percentage to the First Generation. Jour. Heredity 11:365—76. Gowen, John W. 1920c Report of Progress on Animal Husbandry Investigations in 1919. Maine Agr. Exp. Sta. Bul., 283:251—84. Gowen, J. W. 1924a Milk Secretion. Williams and Wilkins, Baltimore, Mary— land. Gowen, John W. 1924b Intrauterine Development of the Bovine Fetus in Relation to Milk Yield in Guernsey Cattle. Jour. Dairy Sci. 7: 311-17. 302 Gowen, J. W. 1926 Studies in Milk Secretion. XVII. Transmitting Qualities of Guernsey Sires for Milk Yield, Butterfat Percentage and Butterfat. Maine Agr. Exp. Sta. Bul. 329, 1—48. Gowen, J. W. 1927 Milk Secretion as Influenced by Inheritance. Quart. Rev. Biol., 2:516-31. Gowen, J. W. 1934 The Influence of Inheritance and Enviromnent on the Milk Production and Butterfat Percentage of Jersey Cattle. J. Agr. Res., 49:433—65. Gowen, M. A., and J. W. Gowen. 1922 Studies in Milk Secretion. XVII. Relations Between Milk Yields and Butterfat Percentages of the 7-day and 365-day Tests of Holstein—Friesian Advanced Registry Cattle. Maine Agr. Exp. Sta. Bul. 306. Graves, R. R. 1926 Transmitting Ability of Twenty—three Holstein—Friesian Sires. U. S. Dept. Agr. Bul. 1372, p. 32.. Hammond, John. 1931 Tropical Dairying Problems. Tr0pical Agr. 8:311—15. Hammond, J., and H. G. Sanders. 1923 Some Factors Affecting Milk Yield. Jour. Agr. Sci., 13: 74-419. Hancock, John. 1949 Studies in Monozygotic Cattle Twins. New Zealand Jour. of Sci. 8: Tech., Vol. 31:2. Harris, G. M., J. L. Lush, and E. N. Shultz. 1934 Progress Report on Comparison of Lactation and Yearly Records. J. Dairy Sci. 17:737-42. Hays, F. A. 1919 Inbreeding Animals. Delaware Agr. Exp. Sta. Bul. 123. 303 Hays, W. P. 1926 The Effect of Environmental Temperature on the Per—- centage of Fat in Cow‘s Milk. Jour. Dairy Sci., 92219-35. Hazel, L. N., and Jay L. Lush. 1950 Computing Inbreeding and Relationship Coefficients from Punched Cards. Jour. Heredity 35:301—06. Headley, F. B. 1933 Effect of Season on Fat Test and Milk Production of Dairy Cows. Nevada Agr. Exp. Sta. Bul. 131. Hills, F. B., and E. N. Boland. I913 Segregation of Fat Factors in Milk Production. Proc. Iowa Acad. Sci., 202195—98. Hoel, Paul G. 1948 Introductions to Mathematical Statistics. John Wiley 8: Sons, Inc., New York. Hutt, F. B. 1934 Inherited Lethal Characters in Domestic Animals. The Cornell Veterinarian, 24:1—25. Johannson, 1., and A. Hansson. 1940 Causes of Variation in Milk and Butterfat Yield of Dairy Cows. Jour. Roy. Swedish Acad. Agr. Jahrg 79. Nr. 6—1/2 [original not seen] Jour. Heredity 32:398-400. Jordan, W. H. 1908 The Feeding of Animals. The Macmillan Co., New York, p. 310 [original not seen]. Kelley, M. A. R., and I. w. Rupel. 1937 Relation of Stable Environment to Milk Production. U. 5. Dept. Agr. Tech. Bul. 591:60. Kendrick, J. F. 1942 Standardizing Dairy Herd Improvement Association Records in Proving Sires. BDIM —- 925. B 304 King, Helen Dean. 1918 Studies on Inbreeding, I. The Effects in Inbreeding 0n the Growth and Variability in the Body Weight of the Albino Rat. Jour. Exp. Zool., 26:1—54. King, Helen Dean. 1919 Studies on Inbreeding, IV. A Further Study of the Effects of Inbreeding on the Growth and Variability in the Body Weight of the Albino Rat. Jour. Exp. Zool., 29:71—102. Klein, John W., and T. E. Woodward. 1943 Influence of Length of Dry Period Upon the Quantity of Milk Produced in the Subsequent Lactation. Jour. Dairy Sci. 26:705—13. Laben, R. C., and H. A. Herman. 1950 Genetic Factors Affecting Milk Production in a Selected Holstein—Friesian Herd. Missouri Agr. Exp. Sta. Res. Bul. 459. LaMaster, J. P., G. W. Brandt, C. C. Brannon, M. H. Fohrman. 1950 Preliminary Report on the Production Records of Cross- bred Dairy Cattle. Jour. Dairy Sci., 33:375—76. Langham, Robert F. 1952 Personal Communications. Lerner, 1. Michael. 1944 Lethal and Sub-lethal Characters in Farm Animals. Jour. Heredity 35:219-24. Linfield, F. B. 1900 Experiments with Dairy Cows. Utah Agr. Exp. Sta. Bul. 68. Linn, J. W., e_t a_1. 1949 Kansas State College Dairy Farm Record Assoc. News Letter [original not seen] Dairy Cattle Breeding. Gil— more, 1952, p. 189. in Ljutikov, K. 1944 Rezuljtaty Rodstvennogo i Nerodstvennogo Razvedenija Bestuzevskogo Skota [The Results of Inbreeding and Out— breeding of Bestugen Cattle] Sovhoz. Proizvod., No. 1, 2:27—36. Anim. Breeding Abstr. 12:191. Lush, Jay L. 1931 The number of daughters necessary to prove a sire. Jour. Dairy Sci., 14.209—20. Lush, J. L. 1940 Intra—sire Correlations on Regression of Offspring on Dam as a Method of Estimating Heritability of Character— istics. Proc. Am. Soc. Anim. Prod.: 293-300. Lush, J. L. 1941 Effects Which Selection of Dams May Have on Sire In— dexes. Jour. Dairy Sci. 24:695—721. Lush, Jay L. 1948 The Genetics of Populations. Mimeographed. Lush, Jay L. 1949 Animal Breeding Plans. Iowa State College Press, Ames, Iowa. Lush, J. L., and A. E. Molln. 1942 Litter Size and Weight and Permanent Characteristics of Ewes. USDA Tech. Bul., 826. Lush, J. L., and R. H. Nelson. 1943 Methods Available for Measuring Transmitting'Ability in Sires and Cows. Jour. Dairy Sci., 262727—28. Lush, J. L., H. W. Norton, and F. Arnold. 1941 Effects Which Selection of Dams May Have on Sire In— dexes. Jour. Dairy Sci., 24:695—721. Lush, J. L., and Floyd Arnold. 1937 Differences Between Records, Real Productivity and Breeding Value of Dairy Cows. Jour. Dairy Sci., 20: 440—41. 306 Lush, J. L., and E. N. Shultz. 1936 Heritability of Butterfat Production in Dairy Cattle. Jour. Dairy Sci., 19:429-420. Lush, J. L., and F. S. Strauss. 1942 The Heritability of Butterfat Production in Dairy Cattle. Jour. Dairy Sci. 252975—98. McCandlish, A. C., L. S. Gillette, and H. H. Kildee. 1919 Influence of Environment and Breeding in Increasing Dairy Production. II. Iowa Agr. Exp. Sta. Bul. 188. Margolin, S., and J. W. Bartlett. 1945 The Influence of Inbreeding Upon the Weight and Size of Dairy Cattle. Jour. Anim. Sci., 4:3—12. Miller, G. D.; J. B. Frye; J. B. Buich, Jr.; P. 1. Henderson; and 1951 L. L. Russof. The Effect of Sprinkling on the Respiration Rate, Body Temperature, Grazing Performance, and Milk Production of Dairy Cattle. Jour. Anim. Sci., 10:961—68. Miller, K., and L. O. Gilmore. 1949 Calf Mortality, Sex Ratio, and Incidence of Twinning in Two University of Minnesota Herds. Jour. Dairy Sci. 32:706—07. Morris, H. P., L. S. Palmer, and Cornella Kennedy. 1933 Fundamental Food Requirements for the Growth of the Rat. Minnesota Agr. Exp. Sta. Tech. Bul. 92. Nelson, R. H. 1941 The Influence of Heredity on the Market Score of Duroc Jersey Hogs. M. S. Thesis. Oklahoma Agr. College Library, Stillwater, Oklahoma. Nelson, R. H., and J. L. Lush 1950 The Effects of Mild Inbreeding on a Herd of Holstein— Friesian Cattle. Jour. Dairy Sci., 33:186—93. 307 Nielsen, J. 1942 Undersgelser over Forekomsten av arvelig Lamhed inden for R.D.M. [Investigations on Hereditary Lameness in the Red Danish Breed]. Nord. JordbrForskn., 24:97—100. Nielsen, J. 1943 Tagttagelser over indon has Hydyr. Nord. Tardhrugsforskning, 5 and 6:204-17, Animal Breeding Abstr., l4:198,224 — 1946. Oloufa, Mohamed M., and I. R. Jones. 1948 The Relation Between the Month of Calving and Yearly Butterfat Production. Jour. Dairy Sci., 31:1029—31. Olson, Thomas N., and George C. Biggars. 1922 Influence of Purebred Sires. South Dakota Agr. Exp. Sta. Bul. 198. Parker, Joseph B., and Charles A. Matthews. 1950 Inheritance of Butterfat Test in the Beltsville Holstein Herd. Jour. Dairy Sci., 33:376. Patow, C. von. 1926 Milchvererbung beim Rind. Sammelreferat. Ztschr. Tierzucht., 61297—354; 529-604. Patow, C. von. 1930 Weitere Studien iiber die Vererbung der Milchleistung beim Rind. Zeit. f. Ziichtung, Reihe B., 21:11-47. Pearl, Raymond, John W. Gowen, and John Rice Miner. 1919 Studies in Milk Secretion. VIII. Transmitting Qualities of Jersey Sires for Milk Yield, Butterfat Percentage and Butterfat. Maine Agr. Exp. Sta. Bul., 281289—164. Plum, Mogens. 1933 Production in a Large Jersey Herd as Affected by Sires, Dam and Yearly Variations. Proc. Amer. Soc. Anim. Prod., 26:53—57. 308 len, Mogens. I935 Causes of Differences in Butterfat Production of Cows in Iowa Cow Testing Association. Jour. Dairy Sci., 18:811- 25. Prasad, Ram Baron. 1951 Ragsdale, 1922a Ragsdale , 1922b Ragsdale, 1948 Ragsdale, 1950 Ragsdale, 1951 Ragsdale, 1949 Estimates of the Heritability of Grease Fleece Weight and Grade and the Repeatability of Grease Fleece Weight in Sheep. Ph.D. Thesis. Michigan State College, East Lansing, Michigan. A. C., and S. Brody. The Effect of Temperature on the Percentage of Fat in Milk. A First Report. Jour. Dairy Sci., 52212—15. A. C., and S. Brody. Seasonal Variations in Percentage of Fat in Cow's Milk. Jour. Dairy Sci., 5:544-54. A. C., Samuel Brody, H. T. Thompson, and D. M. Worstell. Environmental Physiology. II. Influence of Temperature 50° to 105° F., on Milk Production and Feed Consumption in Dairy Cattle. Univ. of Missouri Res. Bul. 425. A. C., H. J. Thompson, D. M. Worstell, and Samuel Brody. Enviromnental Physiology. II. Milk Producation and Feed and Water Consumption Responses of Brahman, Jersey, and Holstein Cows to Changes in Temperature, 50° to 105° F. and 50° to 108° F. Univ. of Missouri Res. Bul. 460. A. C., H. J. Thompson, D. M. Worstell, and Samuel Brody. Environmental Physiology. XII. Influence of Increasing of Temperature, 40° to 105° F. on Milk Production in Brown Swiss cows, and on Feed and Water Consumption and Body Weight in Brown Swiss and Brahman Cows and Heifers. Univ. of MiSSOuri. Res. Bul. 471. A. C., D. M. Worstell, H. J. Thompson, and Samuel Brody. Environmental Physiology. VI. Influence of Temperature, 50° to 0° F. and 50° to 95° F. on Milk Production, Feed and Water consumption and Body Weight in Jersey and Holstein Cows. Univ. of MiSSOuri Res. Bul. 449. 309 Ralston, N. P., S. W. Mead, and W. M. Regan. 1948 Preliminary Results from Crossing Two Inbred Lines of Holsteins on Growth and Milk Production. Jour. Dairy Sci., 31:657. Rasmussen, J. K. 1943 Reed, 0. 1945 Reed, 0. 1946 Reed, 0. 1947 Reed, 0. 1948 Reed, 0. 1949 Reed, 0. 1950 Reed. 0. 1951 Reed, 0. 1946 Arveliy Lambed Indenfor R. D. M. hord. Jordhr Forsku: 3/4:I44-——147 [Hereditary Lameness in the Red Danish Breed] With Reply by J. Nielsen. Abstr., Anim. Breed— ing Abstr., l3—l4z224. E. Report of the Bureau of Dairy Industry. U. S. Dept. Agr. P., 10—12. E Report of the Bureau of Dairy Industry, U. S. Dept. Agr., P., 10—12. E. Report of the Bureau of Dairy Industry. U. S. Dept. Agr. P., 10—12. E. Report of the Bureau of Dairy Industry. U. S. Dept. Agr. P., 10—12. E. Report of the Bureau of Dairy Industry. U. S. Dept. Agr. P., 10—12. E. Report of the Bureau of Dairy Industry. U. S. Dept. Agr. P., 10-12. E Report of the Bureau of Dairy Industry. U. S. Dept. Agr. P., 10—12. E., and M. H. Fohrman. Red Danish Cattle in America. B. D. I. M. - Inf. 31. May, P., I-23. 310 Regan, W. M., and G. A. Richardson. 1938 Reactions of the Dairy Cow to Changes in Environmental Temperature. Jour. Dairy Sci., 21:73—79. Regan, W. M., and G. A. Richardson. 1939 Some Like It Hot -— Some Like It Cold. The Ayrshire Digest, February. Rhoad, Albe rt 0 . 1935 The Dairy Cow in the Tropics. Amer. Soc. Anim. Prod. Proc. 28:212—214. Rhoad, A. O. 1936 The Influence of Environmental Temperature on the Res— ‘piratory Rhythm of Dairy Cattle in the Tropics. Jour. Agr. Sci., 26:36-44. Rhoad, A. O. 1943 The Breeding of Dairy Cattle. Anim. Breeding Abstr., 16:27. Rice, V. A. 194 A New Method for Indexing Dairy Bulls. Jour. Dairy Sci., 27:921—36. Rice, V. A., and F. N. Andrews. 1951 Breeding and Improvement of Farm Animals. Ed. 4, 604- 06. Published by McGraw—Hill Book Co., Inc., New York. Riek, R. F., and D. H. K. Lee. 1948 Reactions to Hot Atmospheres of Jersey Cows in Milk. Jour. Dairy Res., 15:219—26. Rietz, H. L. 1909 On Inheritance in the Production of Butterfat. Rep. of Animal Husbandryrnan. Biometrika, 7:106—26. RobertSOn, Alan 1949 Crossbreeding Experiments With Dairy Cattle. Anim. Breeding Abstr. 17:201—08. 311 Sanders, H. G. 1923 The Shape of the Lactation Curve. Jour. Agr. Sci., 13: 169-79. Sanders, H. G. 1930 The Analysis of the Lactation Curve Into Maximum Yield and Persistency. Jour. Agr. Sci., 201145—85. Schmidt, J. 1948 Schwarzbunte Niederungs—Kiihe und Jersey—Bullen. Kreu— zungsversuche und ihre Auswertung. Ziichtungskunde, 20: 29—39. Seath, D. M., and G. D. Miller. 1947 Effect of Shade and Sprinkling With Water on Summer Comfort of Jersey Cows. Jour. Dairy Sci., 30:255—61. Shull, G. H. 1948 What is "Heterosis"? Genetics, 332439—46. Sinha, K. C., and F. Minett. 1947 Application of Water to the Body Surface of Buffaloes and its Effect on Milk Production. Jour. Anim. Sci., 6:258—64. Skvorcov, V. A., and A. V. Basov. 1940 Metizacija Simmentalami mestnogo skota na Altae. [Cross— ing Local Cattle in Altai with Simmentals.] Trud. vologodsk. se1.—hoz. Inst., 2:107—118. Smith, A. D. Buchanan, and O. J. Robison. 1931 The Inheritance of Milk Yield. Intern. Dairy Cong., Paper No. 4, 127—40. Smith, A. D. Buchanan, and O. J. Robison. 1933 The Genetics of Cattle. Bibliographia Genetica, 1021—104. Smith, A. D. Buchanan, J. R. Scott, and A. B. Fowler. 1931 The Inheritance of Milk Yield in Ayrshire Cows. Jour. Dairy Res. 1:174-79. 311 Sanders, H. G. 1923 The Shape of the Lactation Curve. Jour. Agr. Sci., 13: 169-79. Sanders, H. G. 1930 The Analysis of the Lactation Curve Into Maximum Yield and Persistency. Jour. Agr. Sci., 201145—85. Schmidt, J. 1948 Schwarzbunte Niederungs—Kiihe und Jersey—Bullen. Kreu- zungsversuche und ihre Auswertung. Ziichtungskunde, 20: 29-39. Seath, D. M., and G. D. Miller. 1947 Effect of Shade and Sprinkling With Water on Summer Comfort of Jersey Cows. Jour. Dairy Sci., 30:255-61. Shull, G. H. 1948 What is "Heterosis"? Genetics, 33:439—416. Sinha, K. C., and F. Minett. 1947 Application of Water to the Body Surface of Buffaloes and its Effect on Milk Production. Jour. Anim. Sci., 6:258—64. Skvorcov, V. A., and A. V. Basov. 1940 Metizacija Simmentalami mestnogo skota na Altae. [Cross— ing Local Cattle in Altai with Simmentals.] Trud. vologodsk. sel.—hoz. Inst., 2:107-118. Smith, A. D. Buchanan, and O. J. Robison. 1931 The Inheritance of Milk Yield. Intern. Dairy Cong., Paper No. 4, 127—40. Smith, A. D. Buchanan, and O. J. Robison. 1933 The Genetics of Cattle. Bibliographia Genetica, 10:1—104. Smith, A. D. Buchanan, J. R. Scott, and A. B. Fowler. 1931 The Inheritance of Milk Yield in Ayrshire Cows. Jour. Dairy Res. 1:174-79. 312 Snedecor, George W. 1950 Statistical Methods. Iowa State College Press, Ames, Iowa. Stewart, H. A. 1945 The Inheritance of Prolificacy in Swine. Jour. Anim. Sci., 4:359—66. Swett, W. W., C. A. Matthews, and M. H. Fohrman. 1949 Effect of Inbreeding on Body Size, Anatomy and Pro— ducing Capacity of Grade Holstein Cows. USDA Tech. Bul. No. 990. Tuff, P. 1948 To nye Letalfaktorer hos Storfe. [Two New Lethal Fac— tors in Cattle.] Skand. Vet Tidskr., 38:375—395. Turner, C. W. 1923 Seasonal Variations in Milk and Fat Production. Jour. Dairy Sci. 6:198—204. Turner, C. W. 1925 A Comparison of Guernsey Sires. Missouri Res. Agr. Exp. Sta. Bul. 79. Turner, C. W. 1927a A Comparison of Guernsey Sires. III. Based Upon the Average Persistency of Fat Secretion During the Lacta— tion of the Daughters. Jour. Dairy Sci., 10:479—500. Turner, C. W. 1927b The Mode of Inheritance of Yearly Butterfat Production, Missouri Agr. Exp. Sta. Res. Bul. 112. Tyler, W. J., A. B. Chapman, and G. E. Dickerson. 1947 Sources of Variation in Birth Weights of Holstein Friesian Calves. Jour. Dairy Sci., 30:483—98. Tyler, W. J., A. B. Chapman, and G. E. Dickerson. 1949 Growth and Production of Inbred and Outbred Holstein— Friesian Cattle. Jour. Dairy Sci. 32:247—55. Tyler, W. 1946 313 J., G. E. Dickerson, and A. B. Chapman. Influence of Inbreeding on Growth and Production of Holstein—Friesian Cattle. Jour. Anim. Sci., 5:390—91. Vainikainen, V., and O. Nikkila. 1946 Valerani, 1951 Verna, J. 1945 Ylikylan karjassa sukusiitaksella saavutetuista tuloksista. [The Results of Inbreeding in Ylikyla Cattle] Suom. Maataloust. Seur. Julk., 62:1—174. L. Alcune osservazioni Sull'incrocio di sostituzione Pezzata Nera—Bruna delle Alpi. [Some Observations on the Up— grading of the Brown Swiss Breed by Continuous Cross- ing with Friesian Bulls.] Zootec. Vet., Milano 6:114—17. H. Degree of Repeatability of Production in the Iowa State College Herd With Respect to Quantity of Milk, Quantity of Fat and Fat Percent. Unpublished Master of Science Thesis. Iowa State College Library, Ames, Iowa. Villegas, Valente. 1939 Ward, A. 1940 Washbon, 1947 Washbon, 1948 Washbon, 1950 Livestock Industries of Indo—China, Cambodia, Siam., and Mayla, Philippine Agr., 27:693-725. H., and J. T. Campbell. The Evaluation of Dairy Sires in New Zealand. Emp. Jour. Exp. Agr., 8:249—58. W. E. Dairy Sire Directory and Line Breeding Guide. Vol. 1:1- 50. Box 791, Bath, New York. W. E. Dairy Sire Directory and Line-Breeding Guide. Vol. 11: 1—56. Box 791, Bath, New York. W. E. Dairy Sire Directory, P., 1—40. Box 791, Bath, New York. 1951 1952 1936 Weaver, 1928 Weaver, 1928 1939 Winters, 1943 Winters, I948 I945 Washbon, W. E. Willham, 314 Dairy Breeding Guide. Key Vol. I:1—76. Box 791, Bath, New York. Washbon, W. E . 1952 Supplement to Dairy Breeding Guide Key Vol., 1:1- 16. Box 791, Bath, New York. Waters, N. F., and W. U. Lambert. Inbreeding in the White Leghorn Fowl. Iowa Res. Bul. 202. E., and C. A. Matthews. The Influence of Temperature and Certain Other Factors Upon the Percentage of Fat in Milk. Iowa Agr. Sta. Res. Bul. 107:158—80. Earl, C. A. Matthews, and H. H. Kildee. Influence of Environment and Breeding in Increasing Dairy Production. III. Iowa Agr. Exp. Sta. Bul. 251. O. S., and W. A. Craft. An Experimental Study of Inbreeding and Outbreeding Swine. Oklahoma Agr. Exp. Sta. Tech. Bul. 7. L. N., R. E. Comstock, R. E. Hodgson, O. M. Kiser, and P. S. Jordon. Experiments With Inbreeding Swine and Sheep. Minnesota Agr. Exp. Sta. Bul. 364. L. N., R. E. Comstock, R. E. Hodgson, O.M. Kiser, and P. S. Jordon. Experiments With Inbreeding Swine and Sheep. Minnesota Agr. Exp. Sta. Bul. 400. Woodward, T . E . Some Studies of Lactation Records. Jour. Dairy Sci., 28:209—18. 315 Woodward, T. E., and R. R. Graves. 1933 Some Results of Inbreeding Grade Guernsey and Grade Holstein—Friesian Cattle. USDA Tech. Bul. 339. Woodward, T. E., and R. R. Graves. 1946 Results of Inbreeding Grade Holstein-Friesian Cattle. USDA Tech. Bul. 927. Wriedt, C. 1930 The Inheritance of Butterfat Percentage in Crosses of Jerseys With Red Danes. Jour. Genetics 22:45—63. Wright, Sewall. 1922 The Effects of Inbreeding and Crossbreeding Guinea Pigs. U. 5. Dept. Agr. Buls. 1090 and 1121. Wright, Sewall. 1931 On the Evaluation of Dairy Sires. Amer. Soc. Anim. Prod., Proc., 71—78. Wylie, C. E. 1925 The Effect of SeaSOn on the Milk and Fat Production of Jersey Cows. Jour. Dairy Sci., 8:127—131. Yapp, W. W. 1924 Transmitting Ability of Dairy Sires. Amer. Soc. Anim. Prod., Proc., 90—92. Yapp, W. W. 1929 Transmitting Ability of Dairy Sires Varies Widely. Illinois Agr. Exp. Sta. Ann'l Rep., 125—28. Yapp, W. W. 1930 Dairy Cattle Breeding Investigations. Illinois Agr. Exp. Sta. Ann'l Rep. Yapp, W. W. 1931 Dairy Cattle Breeding Investigations. Illinois Agr. Exp. Sta. Ann'1_Rep. 316 Yoder, Dorsa M., and Jay L. Lush. 1937 A Genetic History of the Brown Swiss Cattle in the United States. Jour. Heredity, 282154—60. Supplementary Citations Dickerson, G. E., and A. B. Chapman. 1940 Eldridge, 1948 Washbon, 1949 Washbon, l 950 Butterfat Production, Reproduction, Growth and Longevity in Relation to Age at First Calving. Amer. Soc. An. Prod. Proc., 76—81. F. E. Evaluation of Dairy Bulls by Multiple Regression and a Study of Variability of Milk and Butterfat Records. Un- published Ph.D. Thesis. Cornell University, Ithaca, N. Y. W. E., and W. J. Tyler. The Number of Proved Sons Necessary to Evaluate the Transmitting Ability of a Sire. Jour. Dairy Sci., 32:706. W. E., and W. J. Tyler. The Number of Proved Sons Necessary to Evaluate the Transmitting Ability of a Dairy Sire. Jour. Dairy Sci., 33:293—298. APPENDIX z22. 8222.22.22.22 esee seed 2am OZ_Z~_.wUZOU :022mE2ou2: 22026322022:— 4 .HHmHmdnH .4OQM4 2222 0.9222220 02 28.33 22.222 meow 222222 $222.2 22822.2. .2. .H .2272 222.20% 2222 22022222202 2.22.2222 22 2222222222222 pw2v2m2®W2 we 9522222». 2222 -228 222 223222? on O2 2222222220 98 22222222252922 .2222 22222.2..«222 022802222 2223 .222 22.222 22222.23 33.2222 :4 .22 2222232222223 220222222n2o 1.22.2 :0 $2282.22 22022222222222 022222 222 2222 222222 9222222 -2228 2mc22 222 22229222222 on. :23 com. mo 22222 4 .m 2222222222 222.2 3.2222220222222222 8. .2.» 222222 2222222222222: 222 m» 2222a $222223 .20“ 32222202222522 222 mm 22222 $822505 222 «m we may 2.222222223222222 .22. 2222 2223 9222.2. .22 :4 O. Q. Q. 4 M222 2222 222222226222 22.223 .3222 wow 222222? 22.2 M2223 22222222222228 0.2.22 .522 mu 222222. we ..40.Q.m.4 22222 no 22.23222uom 0222 02 222222222 -92 22.2222 2222282222 2222222225222 222.2222- mom 22222 222222 so» 02 2822222225 .4 .2 .Q .Q 222 2222.5 22222222232222 222 202222222222 22222 .222 3222226282282 222222 .22 22222? 22 .o .2 2.072 222222.22 2222 2222222392 222.2222 222222 22222222 2322222 .2 .22022222222mw4 22222220 22w22222Q 2232 223282224 2.22 .2222..2o2m2wo.2 e22 2m22222 22222Q .2 6.22m 22m222mQ 22mm 2222 5222222222 .222 22222o2m2m92 23 2.222.22m 2222 2m222>2 .a $2342 .2 .oZ 222.292 2222 2222222222 -92 222.222 M22222 02 2222222222222 222 0.22m 22222.2 22222Q 2222232922 a 2:222 22222222 322222 2222222 22222. 22.222222 h22m2m22m 2.222222 222.2 2232222 on $222th :4 2.822222 22: QQQ 28 2222223222222 mu $283.2 222222 mom P222222 .222 @2222 9822 $222 .22 A. 3222.2 “222222222222 222.2222 83 2 .02 2222222 20222222 2222222892 222.2222 22222222 22:22 232 .o 222022.922 -m2Mo2 222.2 62222222222202 222222222238 .222 22222222222222 222 22222222 2222222222222mm4 22222220 22m2:2.Q 222222 222822222224 22222 2222 2522222222222 3252222222222: 92 25.832922 ma 2292222285 .222 223222222222 22.2.22 .4 .Q .m .D 22222 2222222 22382 23m 232ch .223: .825 22m22222Q pom 25.20.225.222 .222 2222.2222m2mwm 222 “220222222222ow 32222202222222 9.2.2222 222222 232 .22 .222222mm EsQ 22222 220 ESQ 222.222.322.230 2222.22 EmQ 222222.20 2:222Q 28 22.22282 2322 mom .4 2. E. Q 9:222 222.22 .22 2mg: 22.2: ”222222222222 .m 25222322222222 mo 022222 2222 222222222“ we 22222322 22222 .222 52222222on -222 «222.2 m2 =o22s.22m2bo.2 we 22222223222522 022.2. . 2muADm zo2h4~2hm2oum v-1 2222222222 2522222222222 22223 8222222222 2229222322 22222 2283.? 2222222222222 .223 2222 .222 .2272 22222223 n 2022262 .22 MQQQ 02 2222232222322 M22223 2222262222222 224 .22 222922222292 22.2.22 22222222222222.5222 2239 .22 2.222.222.2222 .222 222.2223 222.222 .2222 82222222 222222 23222222“ mm £22222ch 22.22 £228 .muoniofiéol 02 2.22» 2222.» 3222222222222 222 02» 2.2.2 322% 9.022298% 2.22 322222222: : .oZ 222.2223 22.2222 92222222292 222.222 2.28% 2222 02222222 2222 23222 3.22292 £222 224 .m 2220222222222 22.2 .2223 22222223 22.2st 2232 222.222 32 022222 of 2m 2222222222222 222.2222 .22 222222 38 22222232222223 4 .N 222222222“ 22022222222223 a. «22 322222.. 22222 2222222253 2222“ M222 222222222 222.222 .2222 e222w22w now-o2 2222 m2 22 2222222222 22 .222 own 22222 $220282 22 2222222222 2222222222202 222.2222 22222 22 .m2222222222 a one 0222 322332 2222222222 2222222 e22.20.8222 220222222 .222 22222222 .2022 cod» no 23222 .5222 2.2 .2 9222222 922222 om 202.222 ”222.222 no 922222 om 22222223 22022222292 22222.22 22 22222222 232222.232»: .2222 m222w22m 22228222 2223 .222 92.22 22223 $222,222 2222a 32.222 -222 22m22222Q 2282 2222 222222 .323 20.222222 mmouu 222222 .23 222:2 2922.2 :4 28724822022522 252255 .2 $322.2 M22233 -2222 of 2.2223 22222228 252222 £85985» 9‘33 -2225 22222.2 22.2222 .222 22222222232232 22.22222 .20 3222222222 @222 2.22.2 222228 22st 222222 922.232: 222292224 £52,222 0225220222222 2.2.223 H Em "ammo 62222 2922222222222 2222 U255: 092.2222 >mmoo2m$o222 $232522 222228222222». Unmomgm. 32 2222225. 22222226222222 wan U322; m2wmm 22.2222 282 2622228225. 9 82222222225 U. m. 2. 2?. .8322». 2226322222. 22222.. 22322222222222 .2222 2222.2- 3222222225 222.me om 2.22 22222.222 8.3222 0% 22222.22 22.22.222.22; 2226222222225 222222282. Ham 228.22 2222 222222 Usamr 022.3222 2222: 2.222222 2222 2222222. 22.3mm 22». 3822222622 22222.22 $22222 2.2222222 222222222. £322.22. 22222222222025. 2:222 322222 2.1: Um 222222222322. E2222 $222 U222- 2222.322 2226222222022 2222: 22m 222m? .2222 2282.. 222222.28" 222222222 222. $222 2222222222; 2S: 2222 M22022. >222Bm2m 22.2: 22222.22 M28832? 2.5mm... 2.222 82522222222225. 00222028223022 2222: .26 2222232222. 222222225 222 2 2 2 @3532? 2222 .222ch22.2 «222.22 €22: 2222322222222. 222222226 «222% 222 22262522222 002222222022 22: 22.22522 Soc .8 Eco 2222222222? .25 22232 20 no.2 «2 22222.22 222$ 222$ 2m .8 22mm 2.mm.2m$2.m22 22222222222232 222222 UmEmr 2.22225 222329. 2222 2.2228 222.802.2ch .82 22222 22922202922022. .2222 02222222222 no.2 25mm 222 omfimgizsm 225. 322222.22 22225.2 .022 2.220222.me 23.222 3222 -wmmogmfios. «22222 3252 Um U. m. 2. .222. 2mm2m22 2.222. «32.2 22222222 222222 26222.. 22222. 22262222? 222222 2.228222 22.2: 20280322 22 2292522222222 26.- 2.02.22 02 S222 29922232222022. mm 2.2.2: 22262222? 2.22222 2.2.2822? 222. 22: 23222223. @222... 2m 222622 222 22 2.322 Unzip 2222:. 2.2222 2.8222325 8.5 R 2 22222222222. 22.2: 22222 2.2222222.me $232 $222 29.382320? 222222 222222232 mm m 2.532.. 02.0mm... 2232222322222 2.222222an 222.22 2.222222. 222222 222222 222.08%. mm .26.. 223.322 mow S222 :mgobm 02.0mm... 242222.22 02.0mm... 222222 2.2222322 03%.: 12222.22 98% 2.222252%. 222.22 22222522282222 22222522822. 220222.222 03% 225.22% 222.22 222222.. $228.22 2222282222222. 2222222522 25222 $22.22 92222.2. 22222322022 02 92222.2. 22222222222 2222222222 2222 322222.22 €52 $222 >mmoo2m$o22 Sam» Um 26222232222 32 S222 U. 22. 2. 22». 22 222232223 9 22222.2 222.. $222 222:. 22.2.2222 02 .2252 2222222822. 2822222 2222.225 mow 22522225 m: 2.222201% 222.22 2225222212222 252.02%: 222m >mmoo2maos. I_m._.O2~< 2.22m 222622.22 222. 22.222222222222222 222222 U255: 02222222 22mm 222622 22332222322222 SEEMS 2222 2.222222 2222.223 02 22222 22222822 mflmamm U352- 22222222 Om 52.222202222222262 22.222.522.22. 3559a 222 2222222.. 22220222222222 002222.228. «22222 2222222252222 @8222 02222222229 H2222 222.me 2m 2222222 2222022222- 2222.. 22222.8 232222222 32262222 22226222222222: 7.02.222 255222.282 222222222522 :2 2m 222.o<222m 2.2m 2.262.222. 2.22m 22. m. U. .22? «5222222 U. U. 252222222222. 22222 Umm: 02 2229222222923. 22222222222222.2222 @2228 022%? 2222 Mo .8 Um2222222222 M22222 22222.2 22 222222 Moog 2,2322 UmEmr 022.3222. mm mm? 220.822 222262252 222222228 222222 .36 2222228. @22622222m $222222 .8 $222 d. m. U. 22». 2222222222. 2222232 2222.222 22.2 23:32:22. 222222. 22.22 222222 228.22 22222222222322. 9 «2.222222 om 92.222223 222 Manama 0222222322 2522225922. 2222222222322 $222 22922. on $222 6.. m. U. b. 922 222.2225 222222 2222222... 82222 Hams—22m on 222.228.. 222m 35 2222225 22.22 229222422 222222225 3222.22 .222 mmsmmfibm. $222.22 22.. Sam 2222222222222 .8 22922.2- 22222 22. 222222. 2222222222. . 22222222525222 N263 .2222 $5 292.82.22.22 no.3 $222.22. 252.2? 222222262222 «222222322222 22.222222. Mobmwmfiou 2222.222 2222.222 22 moo 222. 2222322212222 22.2222 Um222m22 2222222 2222 32.22-2225 SEQUW 222 22.22 22952.82. 242.222 $222 mmm2m$22om 02 2522222222522 mega 0022mm? S222 mmmewm 22952222322222. 222 222222552 2.2225. S222 292229.532 2222.22 U922- 2m: 92.2.2222 2932289222222. 222222222 2222222222222 :mmmncm2..2m2.m 3.. Emma US$222? KEEN»? 2222225262.. 2222m222mmm 2m 252222222. «2.2253322 222. 22222 2232222 om $222 M22333? flog ZO._.mm wmamsmw o2. 222294.222 2222222222222. 222222 U222- 2m: 9.222222%. mfioow 2m mm22222m mam. 222.22.. 222222222 22.222222 2.2 3: ca 2222262232222. 22 m 22 Um222m22 $22222: 2m 2222222922222 $22.25: 225 22220222222222 22223222222222 222.2222- Em 2223222253022. .5220 222m 022222.22odm2.2m.22o 222222 2.2222 8. 202. 2m 222.3222 8222225222222 222 $222 $222.22 8.03. 222.222 22222 $82222 02.0mm 2222222225 222222 22222.22 222 82222.. . EXHIBIT B . INSTRUCTIONS FOR FILLING OUT 00w WORK SHEET 1. Please do not write above the starred line. 2. If you have a cow or cows in.your herd for wiich you do not receive a work sheet in either of the two groups of work sheets which you will receive, then use one of the blank sheets found in the last group for these animals. It will be neces- sary for you to indicate her ear-tag number and birth date in the female box, the number of her sire in the sire box and the earatag number of her dam in the dam box. The other information above the starred line will be filled in by us after the sheets are returned to us. If this is a purchased cow indicate at the bottom of the sheet from who purchased. If you receive a work sheet for a COW'Which you have sold please indicate to whom sold at bottom of sheet. 3. Please fill in the columns below the "starred line" as follows: Column 1. Enter here the lactation number corresponding to the information filled out for each line across the page. Example: 1 for first lact. — 2 for second and etc. (for each calf) " 2. Enter the freshening date in the order shown at top of Column. Year first, then month and then date. Example: h7—2-2h. " 3. Enter the age at freshening. This must be shown as year and month. Exa ample hOW'tO calculate freshening age. ' If cow calved h6-2—2S assume the date of birth from above starred line under female is hh-l-23. By substracting the birth date from the calving date the age at freshening can be found. Example: b6~2-25 (calving date) hh-l—23 (birth date of cow) 2-1-2 This is the age at freshening written as 2-1, dropping the day since it's less than 15 days. Another Example: h6-7-10 (calving date) h367¢ll (birth date of cow) 2-Ih29 This is the age at freshening and since the days are over 15 the age would be entered as 2-12. As you notice in subtracting the birth date of cow from the freshening date in the last example, 11 from 10 wouldn’t work therefore a month of 30 days was borrowed from 7 to make the days b0. Therefore ho minus ll 3 29 days. This leaves 6 then as the months,6 minus 7 wouldn‘t work. There- fore it is necessary to borrow 1 year or 12 months from.h6. This makes the month 12 +76 2 18, then 18 minus 7 = 11 months. In subtracting the year dates since a year was borrowed from h6 the procedure would be hS minus h? g 2 yrs. h. Enter the length of time between calves. This should be entered by months. 5. Enter the time milked daily.- Ekample: 2,3 or etc. 6. In this column enter the exact number of days for which” roduction is shown ingcolumns in 7 andg8. This must be accurate or errors will result in =:===========F:4— :” +v r - analysis. - 2 - Column 7. Enter here the production record made by the cow that is nearest 305 days. 1| If a 305 day record is shown in the herd book use it. If you have to get the records from the ”total-to—date" column in your herd book take the sub- total that is 305 days or the nearest one over 305 days. Do not cut a record off below 305 days. If a cow produced 270 days show her record then for 270 days. Be sure this is the same length of record as number of days milked in column six. Use only whole numbers. No fraction of pounds. Any partial records should be shown. The fat production which goes in this column is figured exactly like the milk production for column seven and for the same number of days. Use only whole numbers. No fraction of pounds. The disposal of the cow should be entered by one of the code numbers appearing opposite the reasons of disposal shown below. For example, if a cow died from "hardware" number 15 would be entered. If none of the be— low are appropriate leave column blank. 1. Sold — Commercial dairy purposes 2. Sold — Breeding stock 3. Sold — Poor type b. Sold - Low production 5. Sold — Unthrifty, poorfeeder 6. Sold — Hard milker 7. Sold — Udder physically injured 8. Sold — Sterility - bred h times or more, normal heats, still open. 9. Sold — Sterility, lack of heat periods 10. Sold ~ Sterility, apparently settled, showed signs of heat three months later or more. 11. Sold - Bang(s reactor 12. Sold — T. B. 13. Died — Cause unknown 11;. Died — Milk fever lS. Died — Hardware 16. Died - Difficult calving 10. This column pertains to abnormal conditions during a lactation. A code number is entered here, selected from one of the conditions below. If none are apprOpriate leave blank. 1. Physical injury to cow - serious enough to affect production for several weeks. 2. Sickness — unknown, off feed for several weeks. 3. Milk fever. h. Ketosis or acetonemia - not serious enough to affect production Ketosis or acetonemia - serious enough to affect production. 6. Mastitis - serious enough to affect production. 7. Nastitis - not serious enough to affect production. 8. Abortion — 152 days or less after breeding. 9. Abortion — 152 days or more after breeding. 10. Calving trouble. ll. Retained afterbirth. 12. Loss of one or more quarters. .. . - , .- ‘ ‘ .' 1 : " . . . ... ’ ‘ ‘ , < A a . . . . 1. I‘M u. . . \ . . . _ _ a. -‘ J . ' .. , - pl . ; . .-1 I a l t ., ' . ‘.‘ _ ‘ . . 'V A -- u .. 1“ .¢ ' l I ' J -r . i I ' v I - 4., "I 1"; I . - . ‘ - a . ' ‘ , _ - : ,' ‘,. a. U i ‘ .. . .' ‘ A.- ‘ ‘ 7“ .V . \ I .~ . . I x ." ‘.\ . 1 ‘ . . . y - I I. ' . V '. . v ‘ A 1 u ‘- . ... . _ . .,_‘,_ .u . . I. 'n I. J. . . ' . » ;.«.‘. - ' . b - . 0’ y‘ ' A‘ l a . ‘ ‘ . . .- _ t ‘ -< _ ' . .. . . . . . l . \ 4. ‘..l u . u . » _ o‘ ‘ . - . I . . ‘ ‘ g .. t. . . 4, w 0 ’- g . -3- Sex of calf for each calving is shown here also by code number selected from below. 1. Male 2. Female 3. Twins, male and female . Twins, both male 5. Twins, both female 12" If calf was born alive check this column. Example: (;.4) If calf was born dead check this column. Example: (fiv’). If cow had twins and one was dead and one alive check both columns. If calf was abnormal indicate this by entering one of the code numbers selected from below. If calf had an abnormality not appearing below de— scribe this on line in remarks section. 1. Mummified 2. Paralyzed hind-quarters 3. Born blind b. Raw areas on skin 5. Hairless in Spots 6. Pasterns were flexed, calf couldn't stand — later recovered 7. Nervous, twitched all over when frightened. 8. Real small for age — dwarfed Color of the calf is entered in this column also by code number selected from below. 1. Red 2. Black 3. Red and white spotted h. Black and white spotted S. Roan Enter the number of the sire of the calf. Example: D556 Please leave blank — It is for our use in analyzing the data. Remarks Section: This should be used for any additional information you feel we should have that is not shown in the columns. Please return all work sheets regardless of whether you were able to fill it out completely or not. The information above the starred line is needed by us for cer- tain studies. .3; ROOM USE ONLY ”“17 db '-'.'«.;’f'.~_,4¢» _s‘fi- .mij’ llHlWHHHHI!HIHHHIIWWIHWIHHHIVIWHIHHI