‘ m i . .1 g A LABORATORY sway OF GROWTH RATE 3 _ ' m YOUNG _ I 7 - 'MICROTU'S PENNSYLVAMCUS Thai: for tho Door“ of M. S. MICHIGAN STATE UNIVERSITY Theodore Fred Whifmoyor 1.956 A LABORATORY STUDY OF GROWTH RATE IN YOUNG NICRCTUS EKISYLVAYICUS BY Theodore Fred uhitmoyer AN ABSTRACT Submitted to the College of Science and Arts Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF SCIEECE Department of Zoology 19 56 Approved £;Z&7\. ltya /&43;7$QZ_'- Theodore Fred Whitmoyer AN ABSTRACT To analyse growth of young voles, Kicrotus nennsylvanicus, according to recognized statistical procedures, adult animals were bred in captivity under laboratory conditions. Females were isolated when pregnancy was observed so that the young could be weighed and.measured. A total of 320 young animals representing 59 litters were marked at birth and individual records were maintained from which instantaneous growth rates were computed. Statistical analyses of these rates were made. Some of the young and mothers were subjected to various experimental treatments by exchanging half of the offspring between two litters and imposing two different disturbance levels on each (disturbed every two days or every seven days) by handling in measuring and weighing. Since instantaneous rates of growth were calculated for each individual animal, variability could be used as a statistical tool. Analysis of the data revealed the follow— ing information: (1) No statistical differences were shown to exist between instantaneous growth rates of males and females during the first four weeks of age. (2) Large amounts of variation prevent reliable estimates of age by weight criteria. (3) No distinct growth periods were found during at least the first three weeks of growth when analyzed from Theodore Fred Whitmoyer day-to-day individual instantaneous growth rates. (4) Although animals within any given litter were consistent in their rate of growth, the litter's average rate fluctuated during the first three weeks, and a dif- ference was readily identified between litters for short periods. (5) Growth rates of various litters differ so that a significant difference is shown to exist from week to week, but for the over all period no differences between litters were shown. (6) Influencing factors other than heredity, degree of disturbance, parent history, size of litter, and month of birth affect growth rates of young Eggggtgg. These factors may represent climatic factors of the immediate environment, but were not identified in this study. Sex recognition after the first week by means of ex- ternal linear measurements was statistically validated. Other external measurements revealed no differences between sexes for length of tail or hind foot to three weeks of age at which time 51 per cent of the average adult tail length and 89 per cent of the average adult hind foot length was attained. Ear pinnae of all animals unfolded by the fifth day and all animals ha at least one eye Open by the eleventh day. The study poses several interesting problems with regard to growth rate and heredity, environment, and speciation. l“. A LABORATORY STUD‘ CF GROWTH RATE IE YCUHG O Q MIC OTU~ PEYYSYLVATICUS .11 ( By Theodore Fred Whitmoyer Submitted to the College of Science and Arts Kichigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of EASTER OF SCIENCE Department of Zoology 19 56 AC KNOWLEDGE-3133 he author expresses his sincere appreciation to Dr. Don W. Kayne of the Department of Zoology for suggesting the problem and for his tireless efforts in guiding the work of this study. The personal interest of Dr. Karl A. Stiles, Head of the Department of Zoology, is greatfully acknowledged. Miss Gertrude Podsiadly devoted many long hours to the recording of the data. Acknowledgement is made for permission to use the facilities of the Experimental Animal Laboratory of the Department of Zoology; and for certain support from the Michigan Agricultural Experiment Station; and for tenure as Graduate Teaching Assistant in the Department of Zoology. PART III. IV. V. VI. VII. 3L3 OF CCUTEKT roblem and Objectives eview of Literature (33> *4 "U'Ti KATERIALS AID HETIODS A. Breeding Stock B. Laboratory L thods C. Identification and Eeasurements D Experimental Treatments AKALYSIS OF “ESULTS A. Weight gs Age Growth Curves B. Instantaneous Rate of Growth C. Experimental Treatments and Rate of Growth D. Linear Measur ments E. wuailt tive Observat ons DISCUSSIOE SU‘m-‘hUV AID CONCLUsIOIS L TERATVTE CITED APPENDIX 1 ‘3 J p. '1 (4 [1' qum m HH H \‘; I. IKlRODUCTION A. Problem and Objectives Interesting accounts of growth and development of several microtine species have heretofore been published, but with little use of statistical devices to express the variability observed. In the present study, animals were bred in the laboratory to obtain information concerning at least the minimum amount of variability which might be found in such characteristics as change in weight, external linear measurements, secondary sexual characters, and also that variability of growth resulting from the disturbance encountered in obtaining the data. It is eXpected that greater variability may occur in natural populations. ,E. Review 9f Literature In literature dealing with he reproduction of small rodents there is little use of statistical tools in studying variability of growth rate within a taxonomic unit. Gates (1925), Svihla (1932, 1934, and 1936), and McCabe (1950), all working with Peromyscus, point out interSpecific and intersubspecific differences in growth rates, weight, and litter size, but say little about variation within species or subspecies discussed. Among others, Parkes (1926) and Retzloff (1939) have studied growth of the white mouse relative to litter size, and Rabasa (1952) reported varia- tion in growth rate in older white rats in relation to number per cage and temperature. Morrison, £3 a1. (1954) has tabulated a sketchy comparison between several species of Peromyscus and Hicrotus, as well as a species of Clethrionomys and Dicrostonyx. Studies of several species of hicrotus can be character- ized according to content of (a) population dynamics -- Leslie and Ransom (1940), (b) laboratory techniques and studies -- Baker (1932), Pailey (1949), Harrington (1955), Greenwald (1956), Ecke and Kinney (1956), and Ransom (1934), and (0) field or natural history studies -- Bailey (1924), Bodenheimer (1949), Cowan and Arsenault (1954), Hatt (1930), Hamilton (1937 and 1941), Hatfield (1935), Howell (1924), Goin (1943), Barbehenn (1955), and Jameson (1947). Of these papers, anatomical variation is dealt with mainly by Howell (1924) and Goin (1943). In literature of the genus Kicrotus, complete daily growth data is found only in Hamilton's (1937 and 1941) work on E, pennsylvanicus, that of Cowan and.Arsenau1t (1954) on u, qzeggni, and that of Hatfield (1935) on.g. californicus, although no thorough statistical examination of variability is made in any of these. Barbehenn (1955) has shown that "differential growth and size of individuals is not readily attributed to soils, weather, or forage composition" in the field, but no information is given showing variability in daily growth rates, nor do Howell (1924) and Goin (1943) include these valuable data. It remains that only Cowan and Arsenault (1954) have made use of instantaneous rates which may so easily be used for comparing relative growth in different popula- tions. Brody's monumental work (1945) reviews the theory behind instantaneous rates of growth (Chapter 16), and an interesting discussion of the more elementary mathematical concepts involved may be found in Granville (1946, pp. 65, “d 168—171). A simplification of Brody's formula (page an2 - ln.W 508) k - l for use with electric calculator t2"”91 and slide rule may be expressed by the mathematical equivalent: W2 k — ins—,1) . t where k = daily instantaneous rate of growth. Wé fir = ratio of the surviving proportion of weights. '1 t - time interval in days between observations. For invaluable aid in the use of othc r statistical concepts, reference is made to Ostle (1954) and Snedecor (1946). Burt (1954) is used here as the author ty for taxonor ic identification of the Species of: [Sicrotus found in Xichigan; the same book contains a valuable discussion of linear measuring techniques (page 34). In more recent papers, Greenwald (1956) has studied reproduction of Hicrotus californicus, and Ecke and Kinney (1956) have sought to age Microtus pennsylvanicus from seventeen to sixty days of age by using age-molt correlation. 1:— Harrington (1955) has criticized the use of external linear measurements in aging young Iicrotus, and suggests tha bone clearing and staining techniques are valid for deter- mining the age, the method being reasonably accurate to the day for the first two weeks, to the week for the first two months, and to the month for the third and fourth months. This latter method of course, is, as yet, of no use in field investigation. II. MATERIALS ANDZMETHODS A. Breeding Stock Wild voles, males and pregnant females, were brought to the laboratory from a field study at East Lansing, Michigan, during September and October, 1955. Most of these animals previously had been captured and.marked by toe clipping. Offspring of the earlier litters from the wild animals were also used for breeding. Matings were made between available females and several males chosen for signs of sexual activity, using up to eight females with one male. This schedule produced numerous litters, but does not allow ready analysis of genetical factors of growth. ,s. Laboratory Methods Food and water were continuously available to the captive animals. Water was administered by the usual an- nealed—end glass tube and inverted bottle, the distance of the tube end from the cage litter being varied as the young animals grew. Fresh food was administered daily and included corn and rolled cats with fresh carrots and quantities of lettuce trimmings. This lettuce was consumed in quantity by the mice, and was used not only for food and water, but also as nesting material. While long grasses seemed to be preferred for nesting material during the spring, the greater number of litters were raised during the winter in the damp nests of decaying lettuce where rearing success was high. Two types of cages, larger and smaller, were used. Active breeding occurred when five to eight females were placed with o-e adult male in a rather large area. An old aquarium measuring 43 x 20 x 23 inches and several porcelain— ized metal tanks measuring 29.5 x 22 x 15 inches were used for such breeding and for holding excess animals. Animals to be isolated for nursing, for various experiments, and for duration of pregnancy were placed in 12 x 12 x 12 inch gal- vanized metal cages, equipped with removable wire tops. These cages were randomly placed in four tiers of four cages each on movable metal racks. The cage floor was blanketed with two inches of coarse sawdust and fine wood chips through which, to a limited extent, the captive Kicrotus were able to burrow. All animals were maintained in a basement laboratory except for one breeding enclosure which was located in an office for closer observation. The basement laboratory was open to outside ventilation throughout each day. No tempera- ture records were kept; the ordinary temperature was about 70° F. with extremes of perhaps 50° F. to 75° F. The animals were illuminated for at least 16 hours daily, fol- lowing findings of Ransom and Baker (1932). Q. Identification and.Keasurements Individuals of ea h litter were identified shortly after birth by branding one or more quarters with a hot bent dissecting needle. Numbers given at this time ranged from one to eight (see Fig. l), and served only to identify the animal within the litter. It was assumed that no impairment to growth resulted from this treatment. As the fur lenéhened, it was removed with scissors locally, down to the branding scar. Whenever identification was questionable, that animal was not included in subsequent data. After the second week, a number of animals, mostly females, were more permanently marked by toe clipping (see Fig. 2). Notes were made relative to certain qualitative items such as unfolding of ear pinnae, opening of eyes, and other characteristics not discussed here. Linear measurements used were those of tail length, hind foot length, and the distance between the anterior border of the anus and the posterior trunk of the urinary papilla. All mea asurements were made on live animals without anesthetic, using a wooden metric graduated rule and fol- lowing Burt (1954). Animals were weighed on a triple beam balance to the nearest one—tenth gram. While many records were made of the time of measuring, the exact hour of parturition was seldom known, and thus twenty- four hour accm racy in age determination was the basis \a—r of computation. With some departures, most measurements were Figure 1 Diagram to show method of marking infants of a litter by means of a hot bent dissecting needle. OZ) Figure l Figure 2 Method used for numbering older individuals by means of toe clipping. (Feet slightly en arged for detail.) i ’Ilylll I./ ll Figure 2 10 made between 7 P.M. and 1:30 A.E. The litter to be measured was first isolated from the mother and placed in a small pan lined with a folded cloth. The parent was then measured and weighed in a ventilated aluminum cup w'th a spring top and replaced in the cage. Each juvenile was measured and weighed on the open platform of the balance, which was warmed by a loo-watt lamp placed about 10 inches away. After the young were two weeks old it was necessary to confine them in the aluminum cup for weighing. Rate of growth is here discussed in terms of in- U) stantaneous rates, stated on a daily basis, as di cussed by Brody (1945) and used by Cowan and Arsenault (1955). From the original weight data, instantaneous rates of growth were calculated b" simultaneous use of an electric calculator and a log-log slide rule. The ratio of the change of weight was performed on the calculator to three decimal places. The natural logarithm of this quotient was then obtained from the scales of the slide rule and divided by the number of days in the period to put it on a daily basis. The facility of the method as compared to the use of tables re- commends its use to biologists, many of whom may not be familiar with the advantages of the log-log slide rule. The midpoint classes were arbitrarily chosen to al- low the grouping of a number of individual records based upon slightly different times of measurement. In general, each midpoint class contains only these observations which ll assure a strong central tendency about the midpoint. Arbitrary rules were set up to govern the grouping of data, for example, with a growth rate having a midpoint calculated at the sixteenth day 01 age, it would be grouped in midpoint class ten. D. Experimental Treatmen+s To study the possible effects of (a) dividing of a litter between two parents and the inclusion of another animal's young in a litter, (b) the frequent disturbance incidental to measuring the young animals, and (c) the animals nursed by a foster mother compared to those nursed by their own mother, each litter or part of litter raised in captivity was enrolled in one of the following experi- mental categories: (1) Split, disturbed, own mother (2) Split, disturbed, foster mother I (3) Split, not disturbed, own mother (4) Split, not disturbed, foster mother (5) Not Split, disturbed, own mother (6) Hot split, disturbed, foster mother (7) Not split, not disturbed, own mother (8) Hot split, not disturbed, foster mother. (a) The §plit category designates litters which in- cluded a number of animals, usually half of the total litter number, interchanged with another litter. To all appearances, such transfers were readily accepted by the females and were 12 carried out with no more than normal fatality of young. Whenever possible, such transfers were accomplished during the first three days following birth, and between litters born within two days of each other. The Not Split category designates those litters containing only offspring of one parent. When a litter was split, animals were chosen for transfer to a foster parent or retention with the true parent according to coin-tossing or a table-of—random-numbers pro- cedure. (b) Those litters of the Disturbed classification were "disturbed" by handling at least every other day during the first three weeks after birth, and those in the fig; Disturbed category were handled only at seven—day intervals. No doubt all litters were also disturbed to some degree incidental to the daily feeding schedule. (0) Own.Hother versus Foster Mother refers to the parent nursing the animals. "Own" animals were raised by the true parent, and "Foster" animals were raised by a foster parent. III. AEALXSIS OF RESULTS g, Weisht is Age Growth Curves Daily mean weights for both sexes combined are shown in Figure 3, plotted on a linear scale along with one and two standard deviations on either side of the mean. Assuming a normal distribution, the outside curves (two standard }__‘0 mals, while the inside deviations) include 99 per cent of an curves (one standard deviation) include about 66 per cent. In attempting to age a voung animal of known weight, one would read across horizontally from the weight on the ordinate scale, only to find a wide range of corresponding ages. This variability, as observed in the laboratory, I suggests a poor field relationship in judging age from weight, condemning use of weight as more than a general criterion in aging. Although Cowan and Arsenault (1954), indicated a close correlation between weight and age up to thirty-two days after birth for E. oregoni, the present study indicates that the variability reduces the usefulness of this fact for LI . pennsylvani cus . Daily weights of males and females have been combined above because statistical examination failed to reveal sub- stantial differences. he Fisher "t" test, comparing the daily mean weights between males and females, shows no statistically significant difference (based upon the 5 per cent Figure 3 Kean weights of males and females combined (circles) with one and two standard deviations on each sile ot the mean, plotted on a lin ar scale. 11+ 35 +25 30 25 a) -S 2 3‘; .920 3 -28 L- - . :1: 2I5 m 3 IO IO l2 l4 l6 I8 20 22 24 26 28 DAYS AFTER BIRTH Figure 3 15 level of significance) up to the twenty-eighth day, exclud- ing the twenty-sixth day, when for reasons not clear, the difference is very highly significant (based on the 1/10 of 1 per cent level of significance). In general, the lumping of data from both sexes for an overall daily mean for the first three weeks of age is validated by this test. p. Instantaneous Rate 2: Growth The values for individual instantaneous ates of change in weight are shown separately for each sex in F gure 4. One standard deviation is plotted on either side of each mean as an indication of the variability. For none of the time periods shown is there a significant difference between the means of the two sexes. As previously explained, the midpoint classes were set up to accommodate for the fact that not all the calculated growth rates were related to exactly the same periods of time. A cursory examination of this figure suggests that the mean of individual instantaneous rates of growth varies about a value of 0.100 for the first fourteen days, followed by a transition on the six- teenth day, after which a gradual reduction results in a value of about 0.015 by the thirty-second day. There is little apparent difference in the variability exhibited by either sex, but there is a suggestion of reduced variability with increase in age. A closer analysis shows that there is statistically a highly significant difference (at the one per cent level Figure # Mean instantaneous rate of growth of 117 males (solid line) and 90 females (broken line) plotted according to mid— point classes, with one standard deviation on either side as an expression of the variability. (Solid bars represent male standard deviations, open bars female standard deviations.) 16 Figure .4 mwdjo ._.z_on_o_2 v. m. N_ __ o. m . m b 0 m v m N _ IHEm awku< m>