.‘1'. g‘fihm 34*"- ..’ "\ gI’vaIu‘ IV 143%“ . 1'1" (35;; V ‘ . ‘Q . 'ér‘" ‘3‘" I', N111 "‘l‘h‘v‘: ~‘c3‘-%:If"M' ' if “1.? ~ °-':'1.:' ' 1 ‘ ‘ 11 A“, .VIW: ' 4 111151;" Io’f 'r 1 . {0‘6 iii '2‘3‘1‘, 1 (fi‘hfi Elbh‘ m“ . . I: I . ‘1‘."‘113 .:';‘Q"‘ '.l '\W ‘14}; |+\ 8401 I51: ‘11-" " M1 5'.” ““9371: . f“- :" ")",'fIII:l‘ . 0" \| “ o1 I ’ fix .. $6,.“ \ Q 1 \i: . ' MI .1::?-;~4,‘s:::1..1,1 13.11.45,“.9- . A: .. 1" _ W1: £"‘:::?f~+:-I'~ ~- 1.1" I . '18" ' W 13? .- jg. ' 'I‘ .II‘ ,1? ‘Ir ' ‘ {‘46 {APR 1’55} 12:. $11.35 ::.-...I I g \b‘)-~}$§\J ‘35:; Is. ‘1 ‘1‘" 1 1 Wm * “‘3' I .. '6? E: I" l {SF-1:“ ' ” ‘-\. 3‘5“ 1 “N: ... fig; $9111; 1:11" .. I ; I ‘x' \ "'19"; '. We ‘- 1 (PI. 11101 A“: . fufn'. aka-I15 -."»+-,. ; 3% I .. ”(IIIH3KII '51 -. I - 111W 1-112.2;1gsx:.«1-;1f :11; ... 2135131 «Wt-1 Ki, ' ‘WWW’ ' ~ I} 5- .1 11 :97 ‘_ f-JI 1:13 Y¢s_ _ -‘p. \. "3‘ ' ‘ 1‘- fl": :7": l\' - x.) | "‘1‘. ’1.“ 1;: ”:\:'}a A, 6“,." Y 3.1;” ':‘~s..1~,:,”! "Nix if; :' '\::1|' I ’ ‘ 1111‘.‘1‘t\ .v 1‘1.- ): 1'4 II ”11,11. ' ("’2' I ‘fiij‘ cylkfikl'z‘f‘k'u '1 1'4) :V‘ 11;. ‘ a I ' ‘1'.- III: 1- {3'13" \‘fi 5'. 3G; -‘ :‘/? It I: ' 'lIrIgI \ £1§J1M1£AI¢J¥ ’ I .1 9‘2“”..1‘1rTUI‘, ‘1‘” i {IrtIlfi' " .H 1 ‘7- . .. - _ I.v .' .‘. - .V—r'v- ‘1‘ ' ~4~ < ' ‘\ z ' 63‘ fl - ‘ 'm - ' .I' -—,"J- "- .‘T' '1‘ ___——_ 4— "; LIA . I', "121:; ' ' ‘ " .3 : ‘ ‘-‘ 1:32 ('35; 1‘1 21¢ fi%% if: '.1 13* l "III'HVN ‘1 ~". ”‘ "v“‘;IIJ‘:‘Qz§‘I .l'.':\>:1 '11'1 '1; " cnIt‘%{:"\\l' .I .:‘JI"I1’1II.1IJ‘§41:~ww'” "£11. .. 1 1 .1 1" a? ‘3‘ .1}.1‘%1.F:\!'N¢| ‘1’ 21:. ~ ' v . 1H.) z.' ' ‘ 1 1I'..' , fifth]; If '. .. ..o , -11 ""1" l H\‘ g '."‘1-“ , ‘ .*"’LJ-‘&_r ‘;1Ia.~.c.‘.‘v' - A 1 it‘s) 2:;‘25‘ '1‘” Y Ei‘ ”£21.: 1.:1. _.\. t1.“ :( ..'“,- in); ' _._‘( ~ 15‘2“ " : "'..;. 11/, " U Ik‘l'yll. A‘ ”1"." 1 I. .' 1"}! I J'1l'!”‘r'::i'llll fif‘w‘zl , 1133.. -'-'-'~, .1 .:‘-11:3 I‘f .:‘ :Q :1' I {I bun-'1} R‘HC‘I‘CI‘ .11.! I '3 3"""1h' .1 ..‘"i . .‘ I ' . 1:" ‘-‘ . M1 1 01¢.—.- ‘v ‘ . I i.-.¢~’>, “- .. +111}? .' « 1-. I".HII ‘II I II 1’W1I 1 m ‘ ’1’ 1.1151» 1} :17. 35132.1 {I - w 11‘ M"‘1N'M-‘z-I Ill/llIll!!!Hill/ll!!!lllllINN/ll!l l LIBRARY Michigan Stave University This is to certify that the thesis entitled SIGMODON HISPIDUS (RODENTIA) IN LOBLOLLY PINE SUCCESSION AS INFLUENCED BY SUPPLEMENTAL FOOD presented by William Reinhold Teska has been accepted towards fulfillment i of the requirements for | Ph- D! degree inigglggsL may Major professor Rollin H. Baker Date (DIM 24:) \q 79 0-7 639 SIGMODON HISPIDUS (RODENTIA) IN LOBLOLLY PINE SUCCESSION AS INFLUENCED BY SUPPLEMENTAL FOOD by William Reinhold Teska A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Zoology 1978 ABSTRACT SIGMODON HISPIDUS (RODENTIA) IN LOBLOLLY PINE SUCCESSION AS INFLUENCED BY SUPPLEMENTAL FOOD by William Reinhold Teska An experiment was designed to evaluate the following hypothesis: Food is a critical factor in the dynamics of Sigmodon hispidus populations particularly during the winter and in suboptimal habitats. Populations of Sigmodon were studied in six South Carolina loblolly pine plantations of two ages. Initially, the two-year-old plantations (planted in 1973) were characterized by optimal habitat and peak densities of Sigmodon. The three-year-old plantations (planted in 1972) were characterized by suboptimal habitat and declining densities. Populations on the three plantations of each age were live trapped from April 1975 through June 1976. Each plantation had two, 1-hectare, live-trapping grids, one to sample a Sigmodon population which received supplemental food, the other to sample a control population. A mixture of sunflower seeds and oats was provided to each food-supplemented population by maintaining 20 feeding stations and by broadcasting 17 kg of mixed seed weekly from 2 May 1975 through June 1976. William R. Teska During the experiment densities of the 1973 control populations were higher, and the breeding seasons were longer than for the 1972 control populations. Populations from the control grids on both plantation ages, however, exemplified demographic patterns typical for Sigmodon in South Carolina. For both plantation ages supplemental food increased population density, lengthened the breeding season, increased average litter size, increased average growth rate, and improved the physical condition of Sigmodon during the winter. Supplemental food decreased trappable longevity. Decreased survivorship on the foodesupplemented grids may have been due to an increased emigration rate in addition to low trappability of Sigmodon on the food grids. The food treatment had a greater effect on Sigmodon populations in the 1972 than in the 1973 plantations. Population densities were higher on the 1972 food grids than on the 1973 food grids especially during the early summer and the winter. In addition, growth rates for a fall cohort were higher on the 1972 than on the 1973 food grids. Natality rates rather than mortality rates appeared to be the primary determinant of Sigmodon densities. Changes in density were closely associated with the rates of reproduction. Natality rates were dependent on the food supply. Reduced breeding during the winter was accompanied by and may have been caused by poor body conditions. The William R. Teska supplemental food maintained the body condition of the cotton rats during the winter, and a significant number of food-supplemented Sigmodon bred during the late winter. Food appears to be limiting for Sigmodon populations, especially in suboptimal habitat and during unfavorable seasons. The food supply, as distinct from cover, is important in determining the Sigmodon densities in older pine plantations. The distribution and abundance of Sigmodon in areas undergoing rapid succession are strongly influenced by the available food. ACKNOWLEDGEMENTS I am indebted to Rollin Baker for his help and advice during my graduate studies. Also, I extend my sincere appreciation to Michael Smith for his continuous encourage- ment and support of this experiment. Richard Hill, John King, Gerhardt Schneider, and John Fitch provided many helpful comments for which I am grateful. I thank the students and staff of the Savannah River Ecology Lab. for their help and cooperation. Special thanks go to Ken McLeod for his dependable help with the field work and to Paul Johns for his efforts in keeping my vehicle running. Computer programs to analyze much of the demographic data were written and generously provided by Walt Conley. I also thank John Pinder for his assistance with the statistical analysis. Esther Iglich, J. Finder and K. McLeod identified many of the plants on my study areas. Jean Coleman drew the figures. For their encouragement I thank Jean and Bill Wilkerson, Jewel Hayes, Roy Nowlin and Elaine Chappel. The Department of Biology, Furman University facilitated the completion of this dissertation. This study was funded by U. S. Department of Energy contract EY-76-C-O9—O819. The Department of Zoology and The Museum, Michigan State University also provided financial support. 11 TABLE OF CONTENTS Page LIST OF TABLES . . . . . . . . . . . . . . . . . . . iv LIST OF FIGURES. . . . . . . . . . . . . . . . . . . vi INTRODUCTION . . . . . . . . . . . . . . . . . . . . 1 MATERIALS AND METHODS. . . . . . . . . . . . . . . . lO Selection of Study Sites. . . . . . . . . . . . 10 Description of Study Sites and the Experimental Design . . . . . . . . . . . . . . 13 Supplemental Food Treatment . . . . . . . . . . 26 Trapping Methods. . . . . . . . . . . . . . . . 27 RESULTS. . . . . . . . . . . . . . . . . . . . . . . 3O POpulation Numbers. . . . . . . . . . . . . . . 3O Survivorship. . . . . . . . . . . . . . . . . . 48 Reproduction. . . . . . . . . . . . . . . . . . 50 Growth Rates and Body Condition . . . . . . . . 56 DISCUSSION . . . . . . . . . . . . . . . . . . . . . 62 SUMMARY. . . . . . . . . . . . . . . . . . . . . . . 78 APPENDIX . . . . . . . . . . . . . . . . . . . . . . 81 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . 89 iii Table 10. 11. LIST OF TABLES Summary of a Sigmodon trapping survey (16 January to 2 March 1975) on loblolly pine plantations planted from 1970 to 197“. . General information about study sites. Major understory vegetation on study sites Major trees and shrubs on study sites. Average loblolly pine heights, February 1976,1977 . . Minimum number of Sigmodon hispidus known to be alive on each grid from April 1975 to June 1976. . . Minimum number of Sigmodon known to be alive on the combined grids from April 1975 to June 1976. . . . F values and the levels of significance from the analysis of variance for the minimum number 36,37 of Sigmodon known to be alive for each month from April 1975 to June 1976 . The percentage of Sigmodon captured relative to the number of Sigmodon known to be alive on the combined grids from May 1975 to May 1976 . . . . . . . . . . . . . . . Minimum survival rates per one month for Sigmodon hispidus on the combined grids from April 1975 to May 1976 . . . . . . . . Monthly percentage of subadult and adult Sigmodon hispidus with scrotal testes on the combined grids from April 1975 to June 1976. iv Page 12 1A 17 18 25 31 35 A7 A9 51 Table l2. 13. 1A. l5. l6. 17. 18. 19. 20. LIST OF TABLES (CONTINUED) Monthly percentage of subadult and adult Sigmodon hispidus which were pregnant on the combined grids from April 1975 to June 1976. . . . . . . . . The number of Sigmodon hispidus litters and the average size as determined by palpation during two seasons . . . . . . . . . Weight gains of a Sigmodon hispidus cohort from October 1975 to January 1976. Average skin- fold thickness of Si modon hispidus caught in February 1976. Average skin- fold thickness of Sigmodon hispidus caught in March 1976. . Minimum number of Peromyscus pglionotus known to be alive on each grid from April 1975 to June 1976. . . . . . . . Minimum number of Peromyscus gossypinus known to be alive on each grid from April 1975 to June 1976. Minimum number of small mammals known to be alive on each grid from April 1975 to June 1976. . . . . . . . . . . .83, 84, 85, 86, Minimum number of Peromyscus gossypinus known to be alive on the combined grids from April 1975 to June 1976. . .i. . . . . . . . . Page 52 55 57 6O 81 82 88 Figure LIST OF FIGURES Location of the loblolly pine plantations studied; those with the same number (1, 2, or 3) were paired and trapped simultaneously (a trapping set). A 1973 loblolly pine plantation (area: 73-1) photographed winter, 1975. A 1972 loblolly pine plantation (area: 72-2) photographed winter, 1975. A 1973 loblolly pine plantation (area: 73-3) photographed winter, 1976. A 1973 loblolly pine plantation (area: 73-3) photographed summer, 1976. The MNKA of Sigmodon hispidus for the combined food and control grids from both the 1973 and 1972 plantations. . . The MNKA of Sigmodon hispidus for the combined food and control grids from the 1973 plantations . . . . . . . . The MNKA of Sigmodon hispidus for the combined food and control grids from the 1972 plantations O O O O O O O O O I O Ratio of the number of captures on the food grids (N ) to the total number of captures on aIl grids (NT) for nine consecutive days in late October 1975 on areas 73-3 and 72-3. . . . . . . vi Page 16 2O 2O 23 23 3A no “2 A5 INTRODUCTION The mechanisms which regulate the densities of rodent populations remain uncertain despite years of research on this subject. This is probably due at least in part to a complicated multitude of interacting factors which may regulate populations (Lidicker, 1973). A multiple-factor approach to the study of population regulation, although ideal, generally is not feasible because multiple factors can seldom be tested experimentally (Krebs and Myers, 197“). An alternate approach, as used in the present study, is to delineate one or two factors, and to investigate these by designing experiments to test their roles in population regulation (see review in Krebs and Myers, 197A). The food supply of rodents is one factor considered by some not to regulate the densities of natural populations (Chitty, 1960; Krebs and DeLong, 1965; Krebs and Myers, 197A). Fleharty and Olson (1969), for example, investigated the food habits of two rodents and reported that food did not limit the number of these rodents because there was no indication of an over utilization of the food resource. Hairston, Smith and Slobodkin (1960) argued that since herbivores rarely have any detrimental effect on the standing crop, they could not be food limited. Measurements 2 of primary productivity support this belief since herbivorous rodents usually consume less than 5 per cent of the annual production (Golley, 1960; Drozdz, 1966). Grodzinski (1971), however, calculated that rodents could consume up to A7 per cent of the food supply during certain years. Under normal conditions Microtus can significantly affect the standing crop and seed crop of preferred grasses (Batzli and Pitelka, 1970). Recent estimates of the metabolic requirements of some rodents indicate that the available food during the winter may be insufficient to meet the energetic demands of reproducing females (Sadleir, Casperson, and Harling, 1973; Collier 33 a1., 1975). Chitty, Pimentel and Krebs (1968) found that herbivorous rodents had an abundant food supply throughout the year, and that declines in pOpulation numbers were not correlated with changes in the food resource. Other field studies have indicated a positive correlation between the food supply and the rodent population densities and reproductive rates (3. 5., Linduska, 19A2; Myllymaki, 1969; Smirin, 1970; Hansson, 1972). Ferns (1976) noted that an optimal quality forage for Microtus agrestis was not necessarily as abundant as a cursory examination of the habitat might indicate. Many studies, which mostly have been conducted in the laboratory, demonstrate that the quantity and quality of an animal's diet can significantly affect the vigor of the animal. Dietary alterations primarily change an animal's body condition and subsequently alter the growth rate (Millar, 1975; Glass, Harrison, and Swerdoff, 1976), 3 reproductive performance (Gunn and Doney, 1973) and behavior (Frankova, 1973) of the animal. Laboratory studies cannot indicate that the natural food resources are limiting (Ashby and Vincent, 1976); instead, these studies clarify the consequences of an insufficient level of nutrition. The exact role of the food supply is not easily determined since energetic demands are difficult to estimate in the field, and the quantity and quality of the actual food supply are unknown for the typically opportunistic rodent (Batzli and Pitelka, 1971). To test the necessity of a quality food supply for growth and maintenance of free-living populations, experi- mental manipulations of the food supply are required. Typically, supplemental food, usually a grain, has been supplied in feeding stations to a population of rodents. Data collected from live trapping such pOpulations have been compared with the data collected from similar populations which have not received supplemental food. Although not all researchers have examined the same species or the same aspects of the pOpulations, several consistencies resulting from the supplemental food have been noted. With supple- mental food the breeding seasons of Clethrionomys glareolus were extended from a few weeks to an entire winter (Watts, 1970; Bujalska, 1975; Andrzejewski, 1975). The breeding season was also extended or began earlier for Apodemus sylvaticus (Flowerdew, 1972; Watts, 1970), Peromyscus maniculatus (Fordham, 1971), P. leucopus in A Michigan (Lackey, 197“), M. ochrogaster (Cole, 1977), and Mus musculus (DeLong, 1967). Bendell (1959), who examined the effects of supplementary food on P. leucopus in Ontario, failed to note any changes in the breeding season. Supplemental food has particularly enhanced the survivorship of juvenile animals; for example, in C. glareolus (Bujalska, 1975), M. sylvaticus (Flowerdew, 1972), M. ochrogaster (Cole, 1977), and P. leucopus (Bendell, 1959; Lackey, 1974). Supplemental feeding lowered the survival rates for M. californicus (Krebs and DeLong, 1965) but did not change the winter survivorship of g. glareolus (Andrzejewski, 1975). A parameter which has been reported to be influenced by supplementary food and which in turn affects reproductive potential is the growth or maturation rate. Growth rates increased when supplementary food was supplied for g. glareolus (Andrzejewski, 1975; Bujalska, 1975), P. leucopus (Bendell, 1959), g. maniculatus (Fordham, 1971), M. musculus (DeLong, 1967), and M. californicus (Krebs and DeLong, 1965). The increased reproductive performance of the food- supplemented pOpulations generally increased the densities‘ of rodents. Densities of the food—supplemented populations were increased for A. sylvaticus (Flowerdew, 1972; Watts, 1970), P. leucopus (Bendell, 1959; Lackey, 197”), P. maniculatus (Fordham, 1971), 3. polionotus (Smith, 1971), M. musculus (DeLong, 1967), g. glareolus (Bujalska, 1975; 5 Andrzejewski, 1975), and M. ochrogaster (Cole, 1977). Watts (1970) studying g. glareolus and Krebs and DeLong (1965) studying M. californicus reported reverse density trends. The trends from these supplemental food studies vary in degree, and simple statement of trends can therefore be somewhat misleading. For example, Watts (1970) reported a longer breeding season for a food-supplemented population of g. glareolus than for a control population. This increase, however, was only 2 to 3 weeks, and the biological signifi- cance of such an increase is questionable. Other problems compound the drawing of conclusions from such data. Most of the studies have had no replications; and when different researchers have repeated the experiment with the same species at different localities, results have not always been consistent. The inconsistent results partly may have been due to differences in the type of supplemental food and in the size of the study area. Also, many experiments have been short term, some have been coupled to the study of cycles, and most have utilized feeding stations which may alter the movements and social structure of the populations. The present experiment was designed to ascertain the relationships of the food supply and the population dynamics of the hispid cotton rat (Sigmodon hispidgs). The role of the food supply in determining the distribution and abundance of Sigmodgg in areas undergoing rapid succession was also investigated. Supplemental food was used as the experimental treatment. 6 Fundamental differences exist between the nutritional strategies of herbivorous and granivorous rodents (Baker, 1971). Granivores are thought to experience food shortages more commonly than herbivores, partly because of the seasonal production of seeds (Hansson, 1971a). Also, estimates of primary production indicate that more biomass is available as food for herbivores than for granivores (Odum, Connell and Davenport, 1962; Zemanek, 1972). Not all rodents are easily categorized as herbivores or granivores. A large number of rodents eat a variety of foods and should be classified as omnivores (Landry, 1970). Although usually considered to be primarily herbivorous (Fleharty and Olson, 1969; Baker, 1971), S. hispidus may consume substantial amounts of insects and seeds during the fall and winter (Gaertner, 1968; Briese, 1973). Sigmodon hispidus probably should be regarded as an omnivorous rodent. Rodents considered to be largely herbivorous, such as M. ochrogaster (Zimmerman, 1965), M. agrestis (Hansson, 1971b), and Q. glareolus (Holisova, 1975), also eat insects or seeds during certain seasons. It is not known if such seasonal shifts in diet are purely due to phenological conditions. Rodents can have very selective food habits (Gill, 1977) and do not necessarily eat the most abundant food in the habitat (Zemanek, 1972). To examine the importance of the food supply in determining successional patterns of S. hispidus, young loblolly pine (Pinus taeda) plantations were selected as study sites. The characteristic forest succession 7 following the planting of trees affects the density and species composition of the small mammal fauna. For example, a shift in the mammalian species composition following logging has been documented in Oregon (Hooven, 1972), New Hampshire (Lovejoy, 1975), and Ontario (Martell and Radvanyi, 1977). A similar change in species composition occurs in old-field succession in the upper coastal plains of South Carolina (Golley gt al., 1965). The planting of pine trees in plantations affects the understory species composition and greatly accelerates succession (Spurr and Barnes, 1973). The latter effect permits major successional events to be observed within a limited time span. Sigmodon hispidus is a common rodent in the dense grass cover of the early stages of succession (Goertz, 196A). During forest succession, once the understory progresses beyond the grass-forb stage and becomes shaded by shrubs and trees, Sigmodon pOpulations decline sharply in density and vigor (Perkins, 1973). For example, in pine plantations on the Georgia piedmont, this decline occurs after the fourth year following planting (Atkeson, 197A). Much of the research on population regulation has been oriented toward population fluctuations and cycles (Krebs and Myers, 197A). Fluctuations and cycles can be difficult to study because fluctuations can be hard to predict, and cycles require long periods of examination. In the present study, I tried to alleviate the problem of studying unexpectedly fluctuating populations by utilizing the 8 predictability of known successional trends. Populations of Sigmodon were studied in pine plantations of two ages, the younger age characterized by optimal habitat and peak densities of Sigmodon, the older age characterized by suboptimal habitat and declining densities. This experiment, therefore examined the effects of supplemental food on populations of S. hispidus occurring in successionally optimal and suboptimal habitats. Because habitat quality was defined by the stage of forest succession, an insight was also gained into the relationships of the food supply, forest succession, and patterns of mammalian succession. The exact causes of the shifts in the mammalian species composition due to vegetative succession have never been tested by critical experimentation and remain poorly understood. In summary, it was the purpose of the present experiment to test the following hypotheses: Central Hypothesis: Food is a critical factor in the dynamics of Sigmodon hispidus pepulations particularly during the winter and in suboptimal habitats. Null Hypotheses Which Were Tested: 1. Population densities of Sigmodon are not affected by the presence of supplemental food. 2. Survival rates for the control and food- supplemented Sigmodon pOpulations are the same . 9 3. The length of the Sigmodon breeding season and the reproductive effort are the same for the control and food-supplemented populations. A. On food—supplemented grids the average litter size of female Sigmodon is the same as on control grids. 5. Growth rates of a Sigmodon cohort are not changed by the addition of supplemental food. 6. During late winter the body condition of adult and subadult Sigmodon is the same for control and food-supplemented pOpulations. All null hypotheses were tested for populations in succes- sionally optimal and suboptimal habitats. In this way the interaction between successionally determined habitat quality and food supplementation was examined. MATERIALS AND METHODS Selection of Study Sites This study was conducted on the U. S. Department of Energy Savannah River Plant (SRP) in Barnwell and Aiken Counties, South Carolina. The habitat on the SRP is typical of the sandhill and upper coastal plain regions of the southeastern U. S. During the study, temperature and rainfall were average. In the spring of 1975 six loblolly pine plantations on the SRP were selected for study; three each had been planted in 1972 and 1973. Each plantation had been prepared for planting by K-G blading (bulldozing), root-raking, and windrowing (piling the logging debris into long rows about two meters high and several meters wide). All of the areas had been logged during the summer and planted with one-year- old seedlings the following spring. The seedlings were spaced every 2.0 m in rows 3.5 m apart. The rows were parallel to the windrows. Initially, to establish the patterns of mammalian succession, the densities of small mammals were determined on loblolly pine plantations planted from 1 to 5 years earlier. On each plantation, live traps were set in the most suitable habitat for S. hispidus until an index of the 10 ll Sigmodon populations was determined. From two to six plantations of each age were sampled. Trapping was concen- trated in the two-, three—, and four-year-old plantations because these were the only ages characterized by a grass— forb understory. The trapping results (Table 1) indicate that on the SRP S. hispidus populations attained peak densities 2 years following planting and afterwards sharply decreased in density. Because of the low number of cotton rats caught in the four— year-old plantations, the two- (1973) and three-year-old (1972) plantations were selected for intensive study. Natur- ally occurring densities during the late winter of 1975 were the criteria for establishing two-year-old plantations as optimal Sigmodon habitat and three-year-old plantations as suboptimal habitat. Potentially, six 1972 and ten 1973 plantations were available for study. Of these, three of each age were eventually selected. With one type of exception, explained below, the individual plantations were selected independently of the Sigmodon densities. Instead, the plantations were chosen using physical and habitat criteria. The plantations had to be large enough to establish two 1-hectare (150 x 65 m) trapping grids at least 150 m apart and at least 15 m from the edge of the plantation. The selection of the plantations was further limited since each grid was to be divided lengthwise by a windrow. The windrow had to be at least 150 m long and separated at least M5 m from a parallel windrow. In addition, all-weather 12 .ooa mo wzfizaofipase coco .mmm some mo wQOHpmpcwHQ so unmade socoemfim mo Lopez: HBpOp one oucfi Aommv magma: own» no hopes: pmmwpma on» wza©fi>fic mp coapmpcmaa mo own Some so momma: amp» mo Lopez: pcopoQMHU on» you mpooppoo Lopes: omumSwUm mne* III 0.0 0 0m m :50H 0.0 0.2 mm 2mm m 20H H.m m.w 0m 0H: 0 mwma ,0.m o.m ma 00m m :3 III 0.0 0 00 m owma mpcwfiz mane mpnwflz omne moLSmeo mpcwfiz gape mCOprpcme nmucmfim 00H\mopzuom0 *Uopm3m0< 00H\mmg:uowo mmmmamwm no Lopezz pmmw .zmma 0» 050a 8090 coucmHQ mcoapMpcmaa mcfia maaoanoa :o Ammma cozmz m on mpmscmw 0H0 mm>psm wcwaqmnp coooemfim a mo unmeasm .H manme 13 vehicular access to the plantations was necessary. Four plantations of each age met these criteria. During March 1975 the small mammals on each plantation were live trapped. The plantation of each age which had the lowest density of Sigmodon was dropped from the study because logistical limitations dictated that only three plantations of each age could be examined. Description of Study Sites and the Experimental Design The location of each plantation is identified by a U. S. Forest Service timber compartment and stand number. This number, my area code, the year of planting, the size, the elevation, and the soil type for each plantation are listed in Table 2. Figure 1 shows the location of each plantation on the SRP. The species of understory vegetation and of shrubs and trees on these plantations are listed in Tables 3 and A, respectively. Even though the 1973 and 1972 plantations differed in age by only one year, the densities of Sigmodon were greater on the 1973 plantations. The unequal densities were corre- lated with the appearance of the plantations. At the start of the experiment, the two-year-old plantations (planted in 1973) resembled typical old-fields; the planted pines had not grown taller than the grass and forbs (Figure 2). By the third year after planting (planted in 1972), the pine seedlings dominated the grass-forb layer and had begun rapid growth. The three-year-old plantations could be characterized as young pine stands rather than as old-fields (Figure 3). 1A .m03pm ucmmmma esp Ca moan: mmmsp mo comm pom coapwzwamov esp ma mcoo moan one .pac: pcmsmwmcme moa>pmm pmmpom .m .0 oamaomdm m on momma whomess ccmpm 0cm quEpQBQEoo hmpEau mnE "mpoz .asma .ppoamnm<* 0cmm zmsvsm a sammmz moa om mama mums m ma ocmm mmsvsm a gamma: mm ma mama mums oa om 02mm hmsvsm a sagas: moa mm mama mums : ma scam pmpcsc om mm mama mums 5 aa Ucmm ozone a mhsnmmcmpo mm am mama aims om ma ccmm 05099 a mpznmmcmho mm mm mama arms oa ma moome aaom AEV Amsv wcauCMam @000 vampm pcoEpmeEoo coaum>mam omam no know mon< noneae mmnEaB .mouam mBSpm psonm coapMELOkca Hmmmcmo .m manna 15 Figure 1. Location of the loblolly pine plantations studied; those with the same number (1, 2 or 3) were paired and trapped simultaneously (a trapping set). 16 GEORGIA o 5 80" km 000% Lab/o/ly Pine Plan fin g "Pix/e . @ /972 I /.973 SMALL MAMMAL STUDY SITES SAVANNAH RIVER PLANT Figure 1. Table 3. Major understory vegetation on study sites. Dyschoriste oblongifolia Opuntia sp. Cassia fasciculata Lonicera sp. Stipulicida setacea Helianthemum sp. Lechea sp. Aster paternus Aster sp. Carduus sp. Erigeron canadensis Erigeron strigosus Eupatorium capillifolium Eupatorium compositifolium Eupatorium recurvans Gnaphalium obtusifolium Haplopappus divaricatus Heterotheca gossypina Heterotheca graminifolia Heterotheca subaxillaris Senecio sp. Solidago odora Solidago sp. Carex sp. Cnidoscolus stimulosus Baptisia perfoliata Clitoria mariana Lespedeza angustifolia Lespedeza cuneata Lespedeza hirta Lespedeza repens Lespedeza sp. Rhynchosia reniformis Tephrosia sp. Hypericum gentianoides Salvia lyrata Smilax sp. Yucca sp. Polypremum procumbens Passiflora incarnata Phytolacca americana Erigonum tomentosum Rubus betulifolius Diodia teres Galium pilosum Agalinis fasciculata Aureolaria pectinata Vitis rotundifolia Andropogon spp. Aristida sp. Cenchrus echinatus Gymnopogon ambiguus Leptoloma cggnatum Panicum sp. Paspalum sp. Pteridium aquilinum Sorghastrum nutans Triplasis americana Table A. Major trees and shrubs on study sites. Carya tomentosa Crataegus flava Diospyros virginiana Ilex sp. Pinus palustris Pinus taeda Prunus angustifolia Prunus serotina Quercus falcata Quercus 1aevis Quercus marilandica Quercus stellata Rhus copallina Rhus radicans mockernut (hickory) hawthorn persimmon holly longleaf pine loblolly pine plum black cherry southern red oak turkey oak blackjack oak post oak winged sumac poison oak 19 Figure 2. A 1973 loblolly pine plantation (area: 73-1) photographed winter, 1975. Figure 3. A 1972 loblolly pine plantation (area: 72-2) photographed winter, 1975. 20 Figure 2. Figure 3. 21 Figures A and 5 further illustrate the rapid growth of the pine trees. Figure A shows a 1973 plantation during the winter of 1976 when the pines were three years old. Figure 5 shows the same area during the summer of 1976. A paired design was used to test the effects of the food treatment. Each plantation had two live-trapping grids, one to sample a pOpulation which received supplemental food the other to sample a control population. The experi- mental and control grids were randomly designated within each plantation. Previous studies have shown that S. hispidus requires areas of dense ground cover (Goertz, 196A; Wiegert and Mayenschein, 1966). In loblolly pine plantations the piles of logging debris, the windrows, are an important source of cover, particularly during winter when the vegetation dies back. M. H. Smith and L. A. Briese (unpublished) have noted that in the winter, Sigmodon activity is concentrated in the windrows. As mentioned above, each grid was divided lengthwise by a windrow. Because the windrows were stable and were similar on all plantations, winter cover was held relatively constant within and between grids throughout the experiment. On each plantation the two grids were carefully selected so that soil type, the slope, the aspect, and the structure of the vegetation were as similar as possible. The habitat structure reflects the site quality, an important variable, influencing both the naturally occurring food and 22 Figure A. A 1973 loblolly pine plantation (area: 73-3) photographed winter, 1976. Figure 5. A 1973 loblolly pine plantation (area: 73-3) photographed summer, 1976. 23 Figure A. .~ 0'" .0 4" an. 0“. Figure 5. 2A cover. Within practical limits, the paired design controlled for this variable in the tests for effects of supplemental food. A comparative index of the canopy cover and of the quality of growing conditions was obtained by measuring the heights of 56 loblolly pines on each grid. The trees were selected by the point-centered quarter method (Cottam and Curtis, 1956) in which 1A points were used. The pines were measured with a 2.5 m pole in February 1976 and 1977. The same points and the same trees, unless a tree died, were measured each time. The mean heights of the trees for each grid and year are listed in Table 5. Despite the basic similarity in heights between the two grids on each plantation, the heights on such pairs of grids were generally significantly different (P < 0.05; "t" test). There were three exceptions (73—1, 1976; 72-3, 1976 and 1977). The tree heights for the two ages were also significantly different. In 1976, for example, the mean tree heights for the 1973 and 1972 plantations were 1.51 and 2.29 m, respectively. Predation was another important independent variable. Since the control and food populations were generally a few hundred meters apart, a predator seemingly had an equal Opportunity to encounter either population first. The presence of mammalian, avian, and reptilian predators was monitored. Mammalian predators were removed since they disturbed the small mammal traps. Plantations planted in 1972 and 1973 were paired, S priori, for purposes of experimental logistics and 25 Table 5. Average loblolly pine heights, February 1976, 1977. Average Height (m) Area Grida 1976 1977 73-1 C 1.61 i 0.53b 2.37 i 0.79 73-1 F 1.71 i 0.63 2.61 i 0.79 72—1 0 20.1 i 0.69 3.05 i 0.8A 72—1 F 2.29 i 0.66 3.23 i 0.87 73.2 c 1.A7 i 0.A7 2.2M i 0.58 73-2 F 1.21 i 0.31 1.86 i 0.u0 72—2 0 2.08 i 0.58 3.02 i 0.65 72-2 F 1.76 i 0.59 2.63 i 0.76 73-3 0 1.6A 1 0.A7 2.50 i 0.66 73-3 F 1.u1 i 0.37 2.31 i 0.u8 72—3 C 2.81 i 0.81 3.86 i 1.08 72—3 F 2.77 i 0.66 3.7A i 1.09 a Symbols: C = Control Grid; F = Food Grid b i one standard deviation Note: On each grid, 56 trees were measured. 26 subsequent statistical analysis; the pairings are indicated by area code (see Table 2) as follows: 73-1 with 72—1, 73—2 with 72-2, and 73-3 with 72-3. The first number of the code is the year of planting, and the second indicates the trapping set. The trapping set denotes which plantations were trapped simultaneously and received the supplemental food treatment on the same day. This pairing was based partly on the soil type (Table 2) and partly on a growing quality index as determined by loblolly pine heights. As shown in Table 5, set 2 had the lowest growing quality index, set 3 had the highest, and set 1 was intermediate. The pairing, in part, controlled for the effects of site quality upon the treatment. It also helped to control for the effects of daily weather upon the trappability of the rodents. Supplemental Food Treatment A mixture by weight of 1/3 unshelled sunflower seeds and 2/3 clipped oats was used as the supplemental food. Commercial laboratory pellets were not used because they are enriched with vitamins and dietary supplements, and any subsequent response of the rodents might have been due solely to these enrichments. The mixed seed was roughly equivalent to food which a cotton rat normally might consume. Free- living Sigmodon have been reported to eat the seeds of wheat (Fleharty and Olson, 1969) and sunflowers (Goertz, 196A). During the present study, seed preference tests were conducted in the laboratory and small field enclosures to determine the most suitable supplemental food. When tested with whole 27 corn, wheat, cracked corn, milo, red millet, white millet, oats, and sunflower seeds, Sigmodon showed a strong preference for sunflower seeds and oats. The experimental populations received two types of supplemental food applications from 2 May 1975 through June 1976, the last month of live trapping. First, to assure an even distribution of supplemental food and to reduce the possible effect of the food on Sigmodon movement patterns and social structure, the mixed seed was broadcast. At weekly intervals, 17 kg of seed was thrown evenly over each experimental grid and along a 5 m buffer strip surrounding each such grid. Prior to broadcasting, the seed was heated at 690 C in wire mesh containers for 1 week to prevent germination. In addition, mixed seed was provided SQ libitum in 20 feeding stations on each experimental grid. The stations were positioned in between trap sites and at least 15 m from the edge of the grid. At monthly intervals each station was moved to a new location at least 10 m from the previous one. The second type of food application guaranteed that supplemental food was always available. Trapping Methods The small mammals on each study area were sampled with 128 galvanized steel drop-door live traps (6.5 x 9 x 28 cm) arranged in a one—hectare grid (150 x 65 m). A large windrow (average width 5 m) divided each grid lengthwise. On both sides of the windrow, in 16 lines oriented perpendic- ular to the windrow and spaced 10 m apart, the traps were set 0, 5, 15, and 30 m from the windrow. At monthly intervals 28 from April 1975 through June 1976, each pOpulation was trapped for three consecutive nights. The food and control grids on the 1972 and 1973 plantations of each trapping set were sampled simultaneously. Each month the captured mammals were weighed, examined for sexual condition and, if not previously captured, marked by toe clipping. In addition, each month all captured Sigmodon were anesthetized with Metofane (methoxyflurane) for close examination. For Sigmodon males position of the testes and color of the cremaster were recorded. Condition of the vulva, size and color of the nipples, number of palpable embryos, and evidence of lactation were noted for each female. All animals were released at the point of capture as soon as possible. Traps were baited with sunflower seeds. Between trapping periods all traps were locked open. After each Sigmodon capture, a clean, newly baited trap was placed at the capture site to reduce the role of social rank in a heterogeneous trap response by cotton rats (Summerlin and Wolfe, 1973). Because of strong odors remaining in the traps after each Neotoma, Blarina and Cryptotis capture, these traps were also replaced. With a few exceptions, such as June 1975, the supplemental food was broadcast 5 days prior to the first day of trapping each month. While trapping, the food stations were not accessible to rodents. Trapping methods varied slightly with the season. From May to September 10 per cent chlorodane powder was sprayed around each trap to protect trapped rodents from ants. 29 During the winter (November through February) cotton was placed in the traps to provide nesting material. While trapping from May through September, the traps were closed during the day and re-opened in the evening to prevent heat mortalities. During the rest of the year, the traps remained set during the day while trapping. To obtain an index of Sigmodon body condition during parts of February and March 1976, the thickness of four skin folds was measured on each Sigmodon weighing over 50 g. The four measurement locations were the stomach, right shoulder, and right and left hip. In February a Mitutoyo dial caliper (model 505-633) was used. In March a Lange skin-fold caliper set to a constant tension of 10 g was used. The effects of the food treatment on the trap response of Sigmodon were tested on areas 73-3 and 72-3, which were trapped for nine consecutive nights in late October 1975. Supplemental food was broadcast 5 days prior to trapping and again midway through the trapping period. At the conclusion of the study, the rodents on each grid were trapped for seven consecutive nights to remove as many resident animals as possible. After the third night of trapping, all captured animals were brought into the laboratory and killed. RESULTS Population Numbers When permanent trap stations are used, direct enumeration yields the best estimate of relative density (Krebs and Myers, 197A). Densities in the present study are expressed as relative densities rather than as absolute densities since no adjustments for possible edge effects have been made (Smith 2E gl., 1975). If these adjustments were made, the densities would be lowered. For each month the minimum number of small mammals which were known to be alive (MNKA; Krebs, 1966) was calculated. The present report concerns the dynamics of only the Sigmodon populations, and Table 6 details the monthly MNKA for the cotton rats on each grid. Tables 17, 18, and 19 in the appendix show the monthly MNKA on each grid for S. polionotus, for S. gossypinus, and for all the other small mammals, respectively. In addition, for S. gossypinus the monthly MNKA on the three control and the three food grids for each plantation age have been combined in Table 20 of the appendix. The effects of the food treatment and succession on the dynamics of the Peromyscus populations and on rodent diversity will be discussed in a subsequent paper. 30 31 .mnapw Amv coo: 0:0 A00 ao:p:oo 0:p 00p0:wam0c :0pp0a 0:» 0:0 .Azamso0:0pH3Eam 00:00:» 0:03 m:oapmasao: :oacz 0000a0:a :0983: 0500 0:p :uaz 0mospv p00 w:aaqup 0:» ma A0985: 0:0000 0:» .00p:0aa 003 :oau0p:0aq 05p :00» 0:» ma M0263: pmaam 0:8 * aaa a0 m0 om mm mm mm mm mm mm om m: am oa oa mumumm om ma ma ma om mm mm mm mm mm om mm mm ma m oumnmm 0: mm mm mm am am am om mm ma ma mm ma m m mumumm ma m m ca 0 a ma ma ca m m 0 a o a onmumm as m: 0: mm mm as mm am mm om ma m m m aa muaumm : a m m m m m a m m m a a m m ouaumm 03a am mm mm am oo oo m: 0: cm mm mm ma m m mnmumm mm ma ma mm om mm ma mm mm mm mm mm mm ma ma oumumm mm mm am sa om mm mm ma 3a oa ma 0 a o m mumumm :a ma aa a m ma mm oa m m m oa aa m m oumnmm am am mm mm om mm am mm m oa m oa m m m muaumm am ma :a ma om mm ma ma ma oa aa ma mm om :a wouaumm 0:30 am: Ha:a< :ohmz .90m .:00 .000 .>oz .uoo .pa0m .w:< mHSh 0:50 am: Hana< 00h< .mmma 003m on mmma aapa< Bog: vanw £000 :0 0>aa0 0: on :3o:x macamma: :oooewam mo A0985: Essa:az .0 manme 32 The MNKA of Sigmodon for the combined food and control grids from both plantation ages is shown in Figure 6. During April 1975, the month before food supplementation began, relative densities were almost the same, about 38 Sigmodon for both the combined food and combined control populations (Table 7). The relative densities of these populations followed roughly similar trends until September and increased to about 100 animals. From September to January the density of the combined food pOpulations increased to 319 Sigmodon. After January the density of the food populations declined until April at which time sufficient juvenile cotton rats entered the populations to increase the density. The density of the food populations dropped in May but increased by June when A8A cotton rats were captured. The control populations reached a peak density of 139 Sigmodon in October, declined in November, and after a small increase in December continued to decline through May. In June, 110 cotton rats were captured on the control grids. From November 1975 to June 1976, the relative densities of the food populations were significantly larger than those of the control populations (see food treatment analysis, Table 8). Although data from successive months were not independent, the analysis is helpful in evaluating the effects of the food treatment on the populations. The monthly changes in the significance levels were due to variability among the food and control populations. 33 Figure 6. The MNKA of Sigmodon hispidus for the combined food and control grids from both the 1973 and 1972 plantations. 3A I972 and |973 Combined 9 2‘: 1 I 1 I I I I I 1 J F I“ 1A I“ .1 I976 l 1 4A NI J J 1A. 8 CI AI E) l975 L l \u. 0 AK. on I, T I! N .l nu \ C l a. .\\3\ n3 .An\. .1 .m x .T (I 1 n null/I. m m 11/ 1 P F 111/ L n. .11: 1 .m D- v: 1 W 1 b l .0. v. Pull/L b F . h _ mw nu nu nu nu mw nu nu nu nu so nu nu 5 3 3 2 2 II II m>_...< 2>>Ozx mumEDz 55.2.2.5. Figure 6. 35 .m0aLw A00 0000 0:0 A00 ao:0:oo 0:0 0000:0am00 :0pu0a 0:p 0:0 a000:0a0 003 :oau0p:0a0 0:» A000 0:» ma 50:85: 0:9 * :0: mom aom mmm 00m mam mmm mmm 05a maa man maa :0 mm 0: 0 H0009 oaa mp an 00 mm maa mma oma mma 00 mm mm 00 0: mm 0 H0009 awm 0ma 0ma uma :na 00a 00a 02a maa N0 00 0b a: 0m 0m 01mm mm mm om mm mm o: 0: m: ow m: om mm om ma m 01mm mmm Nma mma 00 non mma maa m0 m0 0m 5: m: mm 0 :a 01mm 2» am :2 mm mm me am mm mm a: 0: mm mm :m mm gosmn 0:50 002 Ha:0< :0502 .:0m .:00 .000 .>oz .000 .0000 .m5< ma50 0:50 00: Hana< 00:0 .mnma 0:50 00 mwma aaha< Eopm 00a:w 00:a:Eoo 0:0 :0 0>aa0 0: O0 :30:x :o0oEwam mo :09E5: E5EH:HS .u 0an0e pmm.0 000.0 mm.o mm.o 6am.ma Amm.m mmn.: nn0.aa a m x a :oa000000:H 000.0 00.0 m0.a 00.0 00.0 0:.a 00.0 00.0 a Am 0000000 .0805 0000 00.0 00.0 0m0.ma 00.0 0m.0 00.0 00.0 00.0 a A< 0000000 :oa000:0am 00 00< .>oz .000 .0000 .05< 0a50 0:50 002 HHLQ< 00 000500 .mmna «can 00 000a aapa< 8000 :0:oE :000 now 0>aa0 0: 00 :30:: :o0oEwa0 00 00:85: E5EH:HE 0:0 000 00:0a00> mo mamza0:0 0:0 8000 00:00a0a:0am no ma0>0a 0:0 0:0 005H0> m .0 0a00e .ao.o v 0 n o .mo.o v 0 u n .oa.o v m n m z u 00 :00 06000 mm u 00 Aav 00000 mamammo 06am paaam 0“ "000:0a000 000000000050 37 02.0 00.0 :a.0 20.: :a0.n 000.0 000.: a m x < :0a000000:H nam.0H 000.0 00H.Hm :mm.00 00m.00 000.00 000.:H H Am 000000v .0809 0000 0.0 00.a 0a.» 0:0.0a 000.0 00.: 00.0 a A< 0000000 :0a000:0a0 00 00¢ 0:50 002 aa00< :0002 .:00 .:00 .000 00 000500 .a.0.0coov 0 manna 38 If the MNKA for the 1973 and 1972 plantations are separated (Figures 7 and 8, respectively), the population response on each plantation age can be tested. Utilizing a split plot design, two periods (April 1975 to July 1975 and October to February) had a significant interaction between plantation age effect and food treatment effect (Table 8). This indicates that when compared with the comparable control populations, the food-supplemented populations for each plantation age responded differently to the food treatment. During both of these periods, the 1972 food populations had higher relative densities than the 1973 food p0pu1ations. On the 1973 plantations (Table 7; Figure 7) the combined control p0pu1ations initially had 18 more animals than the combined food populations, but by August both had increased to almost the same density. By November there were more animals on the food grids than on the control grids, and the food populations continued to increase in density until January. The food and control populations decreased in density during the late winter. This decline stOpped in April for the food populations and in May for the control populations. In June, 223 and 7A Sigmodon were removed from the food and control grids, respectively. A different pattern occurred on the 1972 plantations (Table 7; Figure 8). Initially there were 21 more animals on the combined food grids than on the combined control grids. The density of the food populations increased to a summer peak in July, decreased during late summer and then 39 Figure 7. The MNKA of Sigmodon hispidus for the combined food and control grids from the 1973 plantations. '1‘. CONTROL 1 J F M A M J I976 A0 I973 Loblolly Pine Plontings I AMJJASOND I975 300 - I50 - I00 r- . nu 5 >34 Z>>OZv_ mam—232 $552222 Figure 7. Al Figure 8. The MNKA of Sigmodon hispidus for the combined food and control grids from the 1972 plantations. A2 JFMAMJ I976 I AMJJASOND I975 m>_l_< Z>>OZ¥ mam—232 5.3.2.222 .. lllllll 5 1:! L 1 1.1 o 0 \ m n xix W m \x c m A / Du Jl/I. e D I .m m ,P// P F iii/I w. J. m m a m A, 2 7! my. _ b _ _ P b m 0 O O O O O 5 O 5 O 5 3 2 2 I. .I Figure 8. A3 increased steadily until January. A continuous winter decline in density ended in June when 261 Sigmodon were captured on the food grids. The control populations essentially increased in density until October. From October until June, these populations decreased in density. In June, 36 Sigmodon were removed from the control grids. The ability to estimate population size by direct enumeration depends on the trap response of the mammals (Hilborn, Redfield and Krebs, 1976). The food treatment affected the trappability of Sigmodon on each of the food- supplemented grids as evidenced by a reduced trap response. In late October the populations on areas 73-3 and 72-3 were trapped for nine consecutive nights with supplemental food broadcast on the food grids five days prior to trapping and again midway through the trapping period. While trapping, the food stations were not accessible to rodents. Food availability had both a short and long term effect on the trappability of S. hispidus. Figure 9 illustrates the short term or immediate effect of the presence of abundant food on trap response. The ratio of the number of captures on food grids over the number of captures on all of the grids of areas 73-3 and 72-3 combined (NF / NT) for the days 5 through 9 was averaged to obtain an expected ratio of 0.31. The trapping ratio for each of the next four days was compared with the expected ratio. The drop in trappability occurring on the food grids was significant (x2 (dayl) = 20.77, x2 (day2) = 11.u9, x2 (day3) = 8.uu, 2 x (dayu) = 8.67; df = 1; P < 0.005). Figure 9. AA Ratio of the number of captures on the food grids (NF) to the total number of captures on all grids (NT) for nine consecutive days in late October 1975 on areas 73-3 and 72-3. Note: The means of the ratios for each set of grids are indicated by solid circles and the range by the vertical lines. A5 2070403014 0000 002.0 0>40 \u 0 .0 0050nm 00:34 000.0 _ _ 46 On areas 73-3 and 72-3 the MNKA of Sigmodon during October was ll“ and 97 on the food and control grids, respectively. For days 5 through 9, an average of 24 per cent of the known population on the food grids was captured, whereas an average of 6“ per cent of the control populations was captured. The continued low trap response five days after the immediately preceding food application is noteworthy. Most of the broadcast seeds should have disappeared from the ground since rodents, birds, and insects also compete for the grain (Smith and Blessing, 1969). Because cotton rats do not hoard food (Dewsbury, 1970), the long term effect of the supplemental food on the trap response probably was due to other factors such as a change in body condition of the rats. Monthly differences in the trappability of the known population further indicate a reduced trap response for supplementally fed animals. The percentage of Sigmodon captured relative to the number of Sigmodon known to be alive each month was calculated (Table 9). The first and last months of the study were dropped since calculations of trappability are valid only from the second to the penultimate trapping sessions. With the exception of the 1972 grids in May 1975, animals in the control populations were more trappable than those in the populations receiving supplemental food. The high capture rate during October partly was due to trapping areas 73-3 and 72-3 for nine days rather than the usual three days. “7 Table 9. The percentage of Sigmodon captured relative to the number of Sigmodon known to be alive on the combined grids from May 1975 to May 1976. Total Total Month 73-C 73-F* 72-C 72-F C F May 91 33 O 35 63 35 June 80 “3 93 8O 85 67 July 89 37 100 88 9M 70 Aug. 87 32 100 M5 9“ “0 Sept. 91 N6 91 66 91 57 Oct. 95 83 88 77 92 79 Nov. 81 66 91 5” 85 59 Dec. 91 59 89 6O 91 59 Jan. 92 65 85 69 89 67 Feb. 88 71 89 61 88 65 March 94 7O 95 73 9’4 72 April 86 84 87 6M 86 75 May 8M 73 100 52 89 63 * The number is the year the plantation was planted, and the letter designates the control (C) and food (F) grids. MB The reduced trappability of food-supplemented animals is important since it affected the ability to census these populations. The MNKA for the food populations is likely to be a more conservative estimate of the true number of Sigmodon present than the MNKA for the control populations. Because the differences in the trappability would tend to reduce differences measured between the food and control grids, the data are sufficient to reject the null hypothesis that p0pu1ation densities of Sigmodon are not affected by the presence of supplemental food. Supplemental food increased the densities of S. hispidus. Survivorship Minimal calculated survival rates indicate the proportions of the trappable p0pu1ation which when first captured will still be on the trapping grid during the next trapping session (Krebs, 1966). The minimal survival rates for Sigmodon were calculated for each trapping interval of one month (Table 10). Survival rates for the food and control populations were greatest in the winter. For all months except November, the combined control populations had greater survival rates than the combined food populations. The consistency of this trend was significant (Sign test; P < 0.01). There were no significant differences among the control populations or among the food populations within each plantation age. The degree to which the altered trappability of the food populations affected these rates is unknown. The “9 .mmmmzpcopwo CH ma cocoa comm commoaoh pom othQQMo gonad: one nopoz Asfimv 2m. Amov mm. AOOHV om. AHNV mm. Asaav mm. Assv as. am: Ammmv mm. Awav am. Asmv mm. Amav mm. Amafiv am. Ammv mo. Hfipa< Aomav Hm. Azmv as. AmHHV mm. Aamv ms. A350 ms. Aomv mm. 2088: Ammfiv mo. Azmv mm. AmHHV mo. Ammv mm. A550 :5. Aamv 2m. .pmm Anzmv ow. AHQHV mm. Ammfiv as. Ammv mm. Aamv so. Aomv as. .:ms Ammmv as. AHmHV ms. Aflmav ow. A330 om. Ammv ms. A550 as. .omo Aowfiv ms. Amoav as. hafifiv as. Afizv mm. A800 55. “:00 mm. .>oz Amofiv wm. Ammfiv sp. AmOHV om. Azmv as. Ammv am. Aasv as. .880 Aamv m2. Amwv mm. Ammv Hm. Amzv so. Ammv on. Aoav op. .uamm “Nov mm. Amwv mo. Ammv am. Aomv mm. Ammv mm. Ammv mo. .w:< AMHHV Hm. Awwv Fm. Amsv Hm. Ammv mm. Ammv Hm. Aomv mm. sass Aoov mm. Ammv mm. A030 ma. Awmv 2m. Aomv om. Azmv mm. mesh :3 mm. :13 ms. 33 ma. 83 mm. 30 oo. 83 om. >82 Aomv mm. Awmv mm. Ammv mm. Amv om. Amflv ma. AHmV ma. Hfipa< UOOm HOLDCOO 000% HOLuCOD Coot HOLQCOO nuCOE H8809 msmfi msmfi .wsmfi s8: 0» mama Haga< soon mofipw UmcfiQEoo on» so mzofimmfiz coooewflm pom spooe 620 Log mmumh Hm>fi>psm ESEHCH: .OH mamas 5O null hypothesis that the survival rates for the control and food—supplemented Sigmodon populations are the same is rejected. The data indicate that the supplemental food lowered the survival rate of S. hispidus. Reproduction To compare the breeding intensity of the populations, the percentages of males with scrotal testes and of pregnant females were examined. The average litter sizes determined by palpation were also compared. At least 50 per cent of all the subadult (weighing 60 to 99 g) and adult (weighing at least 100 g) males had scrotal testes except during the colder months, October through March (Table 11). During December and January few cotton rats with scrotal testes were captured, indicating reduced breeding. The combined control populations had a higher percentage of scrotal males than the food populations, except during the winter. This difference was significant during three months (£2 (October 1975) = 6.56, 2 (June 1976) = 7.03; df = 1; P < 0.01 and x2 (May 1976) = x 5.97; df = 1; P < 0.05). During two winter months the combined food p0pu1ations had a significantly higher proportion of scrotal males than the control populations (£2 (February 1976) = 7.11, 52 (March 1976) = 8.02; df = 1; P < 0.01). The percentage of pregnant females (Table 12) is the definitive indicator of breeding intensity. From December 51 .mopmwp Hmpopom spas owocu mo owmpcoosoo on» ma seafloo Uzooom on» paw .pchMo moame pazow 6cm vasomQSm mo hopes: on» ma Cezaoo pmaam map wcaomon comm Loco: "mpoz mo mm oo om ao am ow oa mo mm mm om mess oo oo om ma co om oo o ms om ooa aa so: as mm mm mm mm mm om oa so om ms ma aapg< mm so m a: mm a: m ma mm om 0 mm coho: am os m mm om om 0 aa ma o: 3 mm .oom a ooa o om s oo 0 aa 0 o: 0 mm .:ms m oo m N: m a: s :a 0 mm 0 mm .ooa mm o: am am sm sm 0 ca oa ma mm am .>oz mm am mm mm am sa ma aa mm s so am .ooo so ma os mm oo a so s mm o mm ma .oaom mm ma mm om om oa ow :a ooa m ooa oa .ws< so mm om sm mo sa ooa oa mm m am sa sash ss ma om am mm o mo o oo o ooa ma mess ooa m ooa aa ooa o ooa o I- o ooa o so: mo am om ma so ma om m sa o mo ma aapa< Uoom Hoppcoo Uoom HOLpCOO Uoom Hocaucoo SQCOZ amoos msoa msma .osma mash on msma Haao< Eopu moapw UmCapEoo on» so mopmou ampopom spa; msoaomac coooEmHm pasom com pascmnsm no mwmpcooaoo manpcoz .Ha oaome 52 .pcmcwopo omono mo owmpcoopoo map ma cesaoo ocooom on» com .pswsmo woamEom waspm ocm vasomQSw mo amass: oz» ma :Esaoo pmpam on» wcaomon comm pops: "opoz m: oma mm om m: mo om aa N: as a: mm mess mo as mm mm om mm o m as ma 2: oa so: mm mm om mm mm a: om oa mm ma :3 oa aapa< om ooa 0 mm on mo 0 ma os sm 0 ma zoom: sm mm 0 mm om mo 0 ma mm om o sa .oom 0 mm o om 0 mm 0 ma 0 o: 0 mm .coo s oo o a: ma m: 0 ma 0 om 0 mm .omo ma mm m am oa mm o oa m om o ma .>oz m: oo om sm o: om am ma 0: mm mm oa .ooo so sm so om so ma so aa so ma om oa .oamm mo om om sm ms om m: aa 0: ca mo oa .ws< oz am mm om m: om mm m mm aa mm sa sass so mm am ma ss ma mm o a: a ma ma mesh ma ma mo ma mm s ooa m co m mm sa so: mm ma m ma 0: ca ooa a ma o 0 ma aagaa 000% HOLDCOO 000% HOLuEOO 000% HOLpCOU QDCOE amoos msma msma pcmcwopo opoz scan: monaomac copoEwam waspm pom nanomDSm mo owmucoopoo maspcoz .osma oc3h ou msma Hamo4 Eonm moamw oosaneoo on» so .NH oHnt 53 through March, no pregnant cotton rats were found on any of the control grids. During the fall, the Sigmodon in the 1973 control populations bred longer than those in the 1972 control populations. Breeding in the control populations was restricted to the spring, summer, and early fall. The food populations, especially on the 1972 plantations, had a high percentage of pregnant cotton rats during the early summer, 1975. Breeding continued on the food grids for most of the year but at low levels from November to January. Although no pregnant cotton rats were captured in January, two lactating females were captured on a 1972 food grid in February. On the food grids breeding increased during February when 27 per cent of the subadult and adult females captured on the combined food grids were pregnant. Sigmodon embryos could be palpated only after about the fourteenth day of gestation. Therefore, since some pregnant individuals may not have been counted, these figures are only an index of the reproductive condition and not the absolute number pregnant. Except for the first month of food supplementation and in January when no pregnant cotton rats were captured, the combined food populations always had a higher percentage of pregnant subadult and adult females than the combined control populations. The differences for the combined populations were significant in four months (£2 (June 1975) = 7.51, x2 (February 1976) = 10.99, x2 (March 1976) = 15.21, 32 (May 1976) = 8.57; df = 1; P < o 01). SA The breeding season was longer, and the reproductive effort was higher on the food-supplemented grids than on the control grids. The null hypothesis that the length of the Sigmodon breeding season and the reproductive effort are the same for control and food-supplemented populations is rejected. The average Sigmodon litter sizes from the control and food grids were compared for two periods, the fall, September—October 1975, and the spring, May-June 1976 (Table 13). These periods were selected because they are the times of high reproductive activity for Sigmodon (O'Farrell gt al., 1977). During the fall the average litter sizes for the Sigmodon on the combined control and the combined food grids were M.2 and n.8, respectively. In the spring the average litter sizes were less, “.1 and “.7 for the combined control and food grids, respectively. For both periods the average litter sizes on the food-supplemented grids were significantly larger than on the control grids (F (spring) = 5.7M, df = 1, 123, P < 0.02; and F (fall) = 6.80, df = l, 67, P < 0.02). There was not a significant interaction between the treatment and the plantation age which indicates that the food treatment had the same effect on the p0pu1ations of both plantation ages. The null hypothesis that on the food—supplemented grids the average litter size of female Sigmodon is the same as on the control grids is rejected. 55 Table 13. The number of Sigmodon hispidus litters and the average size as determined by palpation during two seasons. Average Season Area N. Size S. D. September- October 1975 73-0* 15 “.20 1.32 73-F 18 5.00 0.77 72-0 13 “.23 0.83 72-F 25 H.72 1.02 May— June 1976 73-C l6 H.19 0.98 73-F 60 9.73 1.09 72-0 5 H.00 1.00 72-F M6 H.70 0.92 * The number indicates the year the plantation was planted, the letter designates the control (C) and food (F) grids. Note: N. = number of litters; S. D. = one standard deviation. 56 Growth Rates and Body Condition The rate of weight gain has been used as an index of the general condition of a population (Goertz, 1965a; Krebs, Keller, and Tamarin, 1969). In the present study the rate of growth was expressed as the weight gain of a cohort of cotton rats. The weight of each Sigmodon weighing less than 43 g in October was compared to the weight of the same individual if it was still present in January (Table 1“). This metric was chosen because it distinguishes all of the cotton rats which were not yet born or were too small to have been captured during the preceding trapping session. It includes only those individuals born between 8 September and 17 October, a period of high reproductive activity. Sexes were considered separately in the analysis. The males gained an average of H2 and 7M g on the combined control and food grids, respectively. The females gained an average of 33 and 56 g on the combined control and food grids, respectively. For either sex the differences in the weight gains by the food and control animals were highly significant (F (males) = 71.87, df = l, 21, P < 0.0001; and F (females) = 33.42, df = 1, 2”, P < 0.0001). In addition, there was a significant interaction between plantation age and the treatment indicating that the food treatment had a greater effect on animals in the 1972 populations than on those in the 1973 p0pu1ations (F (males) = 11.00, df = 1, 21 P < 0.01; and F (females) = 8.73, df = 1, 24, P < 0.01). 57 Table 19. Weight gains of a Sigmodon hispidus cohort from October 1975 to January 1976. Average Sex Area N. Weight Gain (g) S. D. Males 73-C* 10 99.9 3.69 73-F 6 67.5 10.56 72-C 6 37.3 5.05 72-F 3 87.0 21.79 Females 73-C 7 39.0 6.95 73—F 6 99.2 9.91 72-C 5 31.9 8.08 72-F 10 63.2 11.09 * The number indicates the year the plantation was planted, the letter designates the control (C) and food (F) grids. Note: N. = sample size; S. D. = one standard deviation. 58 Growth rates were higher on the food grids than on the control grids. The null hypothesis that growth rates of a Sigmodon cohort are not changed by the addition of supplemental food is rejected. An indicator of an animal's condition is the amount of subcutaneous fat (Kolodziej-Banach, 1976). An index to the amount of this fat can be obtained by measuring the thickness of a fold of skin (Sloan, 1967). The four skin folds which were measured on each cotton rat weighing over 50 g in February and March were averaged. Data were analyzed separately for each sex and month of sampling. During both months (Tables 15 and 16) the average skin folds for either sex were significantly thicker on the combined food grids than on the combined control grids (F (February males) = 36.78, df 1, 78, P < 0.0001; F (February females) = 52.63, df l, 95, P < 0.0001; F (March males) = 36.95, df 1, 111, P < 0.0001; F (March females) = 38.95, df = l, 137, P < 0.0001). For example, in February the average thicknesses of the male skin folds were 1.12 and 1.37 mm on the control and food grids, respectively. There were no significant interactions between the food treatment and the age of the plantations for either February or March. This indicates that the food treatment had the same effect on the populations of both plantation ages. The body conditions of Sigmodon, as determined by skin—fold measurements, were better on the food grids than on the control grids. The null hypothesis that during late 59 Table 15. Average skin-fold thickness of Sigmodon hispidus caught in February 1976. Average Sex Area N. Thickness (mm) S. Males 73-0* 21 1.19 0.15 73-F 28 1.31 0.20 72-C 10 1.07 0.13 72-F 23 1.95 0.18 Females 73-0 13 1.09 0.15 73-F 26 1.39 0.26 72-C 16 1.07 0.17 72-F 99 1.99 0.19 * The number indicates the year the plantation was planted, the letter designates the control (C) and food (F) grids. Note: N. = sample size; S. D. = one standard deviation. 6O Table 16. Average skin-fold thickness of Sigmodon hispidus caught in March 1976. Average Sex Area N. Thickness (mm) D. Males 73-C* 30 .13 .16 73-F 27 .31 .16 72-C 15 .06 .17 72-F 93 .28 .17 Females 73-C 21 .07 .18 72-0 20 .07 .15 72—F 69 .29 .19 * The number indicates the year the plantation was planted, the letter designates the control (C) and food (F) grids. Note: N. sample size; S. D. one standard deviation. 61 winter the body condition of adult and subadult Sigmodon is the same for control and food-supplemented populations is rejected. DISCUSSION Populations of rodents vary in density through time and space. The densities of rodent populations probably are regulated by a multitude of interacting factors (Lidicker, 1973). The availability of food is just one of the suggested factors (g. g., Kalela, 1962; Schultz, 1969), but there is little experimental data with which to evaluate any of the proposed factors. Because rodents rarely consume more than a small per cent of the annual primary production (3. g., Golley, 1960), they are not thought to be limited by food (Hairston, et al., 1960). These rates of consumption, however, do not take into account that food availability varies with the season and the habitat. During unfavorable seasons and in unfavorable habitats, food is more likely to become limiting. Further- more, the actual food supply may be less than estimates of primary productivity indicate since rodents can be very selective in food habits (Gill, 1977). The productivity estimates also do not account for seasonal variations in food quality. The importance of habitat quality and of seasonal changes in the food supply for the hispid cotton rat was 62 63 investigated in the present study. The central hypothesis was that food is a critical factor in the p0pu1ation dynamics of S. hispidus populations particularly during the winter and in suboptimal habitat. To evaluate the central hypothesis, six null hypotheses were tested. Each of these will be discussed with respect to the general population dynamics of S. hispidus and other rodents. The role of the food supply in determining habitat quality also will be discussed. Of the six testable hypotheses, the first concerns the effects of food on population numbers. The data clearly establish that food limited Sigmodon densities. The maximum densities of Sigmodon previously reported for populations on the SRP ranged from 10 to 37 per hectare (Golley e3 gl., 1965; Schnell, 1968; Gardner, 1975). The highest MNKA for the 1973 and 1972 combined control grids were within this range, 27 and 20 per hectare, respectively. The highest MNKA for the 1973 and 1972 combined food grids were 79 and 87 per hectare, respectively. The densities on the combined food grids are greater than any densities reported for the SRP. The supplemental food altered the timing of the seasonal population fluctuations. The combined control grids exemplified patterns typical for Sigmodon in the southeastern U. S. (Layne, 1979) and in South Carolina (Gardner, 1975). When compared with the control populations, the densities of the food populations continued to increase longer in the fall and began to increase earlier in the spring. 69 The increase in density due to the supplemental food was predicted, but not altogether expected since supplemental food experiments with similar rodents have produced equivocal results. For example, Krebs and DeLong (1965) failed to obtain high densities of Microtus with supplemental food. In another supplemental food study (Cole, 1977), the addition of food increased Microtus densities but failed to prevent a general population decline. In both of these experiments the factors which caused the low densities were not determined. The experimental design may have influenced the conclusions since neither study had any replicates, and the food- supplemented areas were less than 1 hectare in size and were smaller than the control areas. In addition, the study areas of Krebs and DeLong were 9 km apart. To account for the natural variability within and between small mammal p0pu1ations, it is essential to replicate experiments. In the present study, the increased densities with food supplementation indicate that food was limiting, but density changes alone cannot clarify the mechanisms which cause the changes in p0pu1ation size. Fluctuations of population density are due to the effects of one or a combination of three factors, emigration-immigration, natality, and mortality. The following discussion and the rest of the null hypotheses concern these factors and their relative influence on Sigmodon densities. Schnell (1968) and Fleharty, Choate, and Mares (1972) concluded that densities of Sigmodon are determined primarily by the mortality rates, especially from predation. Few 65 studies, however, have documented the mortality rates of Sigmodon. Data from enclosed populations led Wiegert (1972) to suggest that avian predators had a larger effect on Sigmodon pOpulations than mammalian predators. In the present study few avian predators were seen, and there was little evidence to suggest that mammalian predators concentrated on the high density populations. In late April 1976, however, someone accidently released a number of raccoons (Procyon lotor) near two of the study sites. While trapping the food grids on area 72-3 in April and on areas 72-3 and 72-2 in May, the traps were heavily disturbed by raccoons. Density estimates for these months were lowered because of the trap disturbances and perhaps because of predation. Whether the raccoons were attracted to the food grids by the high rodent densities or by the food readily available in feeding stations is not known. Although no one has studied the impact of snakes on Sigmodon populations, snakes may have been an important cause of mortality, especially on the food grids. Relatively few snakes were seen during the study, but most were located on the food grids. Twice during June 1976, when Sigmodon were at their highest densities, two snakes (Pituophis melanoleucus and Coluber constrictor) were observed in the afternoon killing Sigmodon. Perhaps related to the predation by snakes, trappable longevity was greatest during the winter when the snakes were not active. 66 Predation certainly is capable of lowering population densities. But, since densities on the food-supplemented grids were twice as high as the "predator limited" densities suggested by Schnell (1968), it can be concluded that other agents were more important in determining Sigmodon densities in the present study. From the second hypothesis, the data tend to indicate that supplemental food lowered the survival rates of Sigmodon. Several factors probably were responsible for the apparent low survival rates of Sigmodon on the food grids, but the major reason may have been the reduced trappability of the Sigmodon on those grids. Hilborn e3 al. (1976) established that low trappability impairs the ability to census popula- tions and to obtain an accurate measure of survivorship. This impairment would be particularly significant with an animal, such as Sigmodon, which has a high turnover rate. Goertz (1969), for example, reported that of the Sigmodon which he trapped, about 50 per cent had disappeared by the second month and 98 per cent by the sixth month. Chances are that if a marked individual was not captured during a particular trapping session, it would disappear before the next trapping session. Survival rates would therefore appear artificially low. Although difficult to distinguish, mortality and emigration are the primary factors responsible for reducing survival rates. The main causes of mortality would have been predation and starvation. Data to be discussed later about 67 Sigmodon body condition suggest that starvation was not the cause of the low survival rates on the food grids. Predation, which has been discussed, also was probably not a significant factor. Increased snake predation on the food grids may have contributed somewhat to the reduced survival rates. During the present study emigration and immigration were not studied intently, but there was no evidence to suggest that large numbers of Sigmodon moved into the food grids from surrounding areas due to the availability of food. The net movement of cotton rats may have been in the opposite direction. Of the 2093 Sigmodon marked during the study, 50 switched grids, a distance of at least 150 m. Of these Sigmodon a significant number, 39, were first marked on a food grid and later caught on a control grid (S2 = 7.89; df = l; P < 0.01). Increasing densities, such as those which occurred on the food grids, have commonly been associated with rodent dispersal (Lidicker, 1975; Krebs et al., 1976). The emigration from the food grids, compounded by the low trappability of Sigmodon on those grids, was most likely responsible for the low survival rates on the food grids. Because the survival rates were influenced by rodent trappability, dispersal, and the food treatment, it may be unwise to draw conclusions about the actual survival rates from these data. Similarly, previously reported effects of supplemental food on rodent survivorship are difficult to evaluate. Each supplemental food study has used different methods of trapping; for example, in several studies the 68 traps were placed within the feeding stations. Their placement probably increased rodent trappability and hence survival rates. The third hypothesis deals with the effects of supple— mental food on natality, specifically on the breeding intensity and the length of the breeding season. The conclusion from these data is that the level of the food supply, particularly during the winter, controlled breeding. Typically Sigmodon in the southern U. S. have a bimodal reproductive cycle with peaks of activity in the spring and fall (Layne, 1979; Bigler and Jenkins, 1975; O'Farrell gt‘al., 1977). During the winter on the SRP, pregnant females rarely have been caught; and males have reduced testicular size (Briese, 1973; O'Farrell 22 al., 1977). Similar seasonal trends occurred on the control grids in the present study. 0n the other hand, the food populations bred throughout most of the year, including the winter when a few pregnant females were captured. The percentage of females which were pregnant when captured indicates that breeding began two months earlier on the food grids than on the control grids in the spring of 1976. From correlative data others have concluded that for Sigmodon the cessation of breeding during the winter was due to unfavorable weather conditions (Goertz, 1965b; Cameron, 1977) along with a decline in the quantity or quality of the food supply (Negus, Gould, and Chipman, 1961; Layne, 1979). 69 The present study provides the first experimental data to substantiate the effects of food on the breeding intensity of Sigmodon. An important aspect of natality is the litter size. Cole (1977) noted that supplemental food increased the litter size of M. ochrogaster, and a similar trend was expected in the present study. The effects of nutrition on the average litter size during two seasons were investigated with the fourth hypothesis. For both seasons the average size of palapable litters was significantly greater on the food grids than on the control grids. It should be noted that the average litter sizes obtained during the present study are not directly comparable with other reported litter sizes, since those sizes generally have been determined by dissecting the animals. For example, O'Farrell 93 21° (1977) found an average litter size of 9.66 for Sigmodon on the SRP. This is larger than the average litter sizes for the control populations, and therefore, it appears that palpations underestimate the actual litter size. Sigmodon litter size varies throughout its geographical range. Data presented here support the probable explanation that this variability represents a response of Sigmodon to local conditions. The litter size of cotton rats is positively correlated with female age (Layne, 1979) and body size (Goertz, 1965b). As will be discussed below, the supplemental food increased growth rates and hence affected the age- weight distribution of the p0pu1ations. Therefore, it is not possible to state conclusively from these data if the 70 increased litter size was directly due to the improved food resource or indirectly to the altered age-weight distribution of the females. In either case, however, the increased litter size on the food-supplemented areas represents an increase in the natality rate, an important factor in population growth. In most previous studies, supplemental food has been reported to increase individual growth rates (g. g., Krebs and DeLong, 1965; Andrzejewski, 1975; Cole, 1977). Findings in the present study were consistent with those reported for other rodents. Data from the test of the fifth hypothesis indicate that growth rates from October to January of a fall-born cohort were significantly greater on the food grids than on the control grids. The age to sexual maturity of rodents is determined in part by individual growth rates (Cole, 1959). In addition, litter size is correlated with female body size, therefore faster growing females may produce larger litters (Goertz, 1965b). In such ways, increased individual growth rates can cause an increase in the population growth rate which subsequently may result in an increase in density. Individuals usually have been reported either to lose weight or at least to gain very little weight during the winter (Goertz, 1965a; Fleharty and Choate, 1973; Layne, 1979). During the late winter of the present study, Sigmodon on the control grids appeared thin and emaciated, especially when compared with robust Sigmodon on the food 71 grids. With the sixth hypothesis, the body conditions during the winter were investigated. One recurrent problem in population biology is to accurately quantify changing body conditions of living rodents. Fleharty, Krause, and Stinnett (1973) found that lipid reserves in cotton rats were depleted during the winter, but this determination required killing the animals. In the present study measurements of the skin—fold thickness on living animals were utilized. These measurements can be used to obtain a reliable index of body fat and condition of living rodents. The average skin-fold thickness was signifi- cantly greater on the food grids than on the control grids. The present data suggest that nutrition may have controlled the breeding intensity by affecting the total body condition of the rodents. These data are not sufficient to establish a causal relationship between body condition and reproductive performance, however, such a relationship exists for laboratory rodents (Millar, 1975). Similarly, the body condition may have affected the reproductive rates in other supplemental food studies in which a grain was supplied. Negus and Berger (1977) recently reported that limited supplements of green forage can start reproduction in nonbreeding winter populations of M. montanus. Chemical stimulants were thought to be the factors which increased breeding intensity. These studies suggest that the role of nutrition in determining reproductive output is complex and can be influenced in a number of different ways. Additional studies need to be conducted to clarify the role of the 72 quantity and quality of the food supply in determining body condition and reproductive output. Throughout the U. 8., winter appears to be the critical season for cotton rats. As previously noted, p0pu1ation numbers decline, breeding rates decrease, and body conditions deteriorate during the winter. Particularly for northern populations, poor body conditions (Dunaway and Kaye, 1969; Fleharty g3 g1., 1972) and population declines (Goertz, 1969) during the winter have been associated with cold temperatures. Although difficult to document in the field, Sigmodon mortality was attributed directly to cold temperatures combined with starvation in at least one case (Dunaway and Kaye, 1961). Since Sigmodon populations decline in vigor during the winter in southern as well as northern p0pu1ations, the simplest explanation for the seasonal decline is shortages of food and not cold temperatures Egg gg. Similar conclusions from previous studies on Sigmodon populations were based on correlative information (Layne, 1979; Fleharty g§.gl., 1973). The present study provides experimental data which clearly indicate that the winter decline was due to an insufficient supply of food. All six of the null hypotheses were tested for p0pu- 1ations of S. hispidus in successionally Optimal and suboptimal habitats. Habitat quality was defined by the stage of forest succession. Previous comparisons of logged areas with mature forests suggested that the shift in species composition of small mammals which accompanies forest succession may be caused by differences in cover and food 73 (Lovejoy, 1975; Martell and Radvanyi, 1977). In addition to the obvious changes in plant life forms, succession affects the nutritive content of the plants. Klein (1965) reported that the nutritive content of shaded plants growing on the forest floor was less than that of plants growing in open areas. The present study is the first experimental manipulation of the food supply of rodents in areas undergoing rapid succession. The evidence suggests that low densities of Sigmodon in older pine plantations were primarily due to inadequate food resources. In comparison with the p0pu1ations on the 1972 control grids, higher densities and longer breeding seasons observed for those on the 1973 control grids indicated that the Sigmodon habitat was more favorable on the 1973 plantations. Yet, the 1972 food grids generally had higher densities of Siggodon than the 1973 food grids. The significant interaction between the effects of the food treatment and the plantation age on density support the importance of the food supply. The significant interaction during the early summer of 1975, partly a function of the sharp increase in densities on the 1972 food grids, is noteworthy. It may indicate that food was more limiting on the 1972 grids than on the 1973 grids even during the summer. The significant interaction between the effects of the food treatment and the plantation age on growth rates also demonstrates that food is more limiting on the older plantations. Higher densities on the combined control grids at the end of the study than at the start, despite unfavorable 79 succession of the vegetation, may have been due to the dispersal of individuals from the food grids. The above discussion regarding forest succession is applicable to a discussion of habitat quality since at the start of the experiment, the 1973 and 1972 plantations were designated optimal and suboptimal habitat, respectively. Habitat quality for rodents is determined by a number of interrelated characteristics, but cover and food are two of the primary components. A common difficulty in evaluating these two factors is that for many rodents the source of cover also provides the source of food. Because of this, only a few researchers have been able to separate the effects of cover and food. For example, food may be more important than cover in determining the distribution of M. californicus populations (Batzli and Pitelka, 1970). Batzli (1979) and Cole (1977) showed strong positive correlations between habitats with an optimal, quality forage and the demographic characteristics of Microtus populations. When supplemental food was provided to M. ochrogaster in a suboptimal habitat (Cole, 1977), the food population increased in density and showed higher reproductive rates and higher average growth rates than the control population. The supplemental food, however, did not prevent declining densities which occurred on the study site. Regretfully, Cole did not simultaneously feed pOpulations living in optimal habitat. 75 For Sigmodon populations adequate cover is a recognized feature of optimal habitat (Goertz, 1969). In the present experiment the windrows were an excellent source of cover. During the winter virtually all Sigmodon captures occurred within about five meters of the windrows. An unique aspect of this experiment was that although the windrows provided abundant winter cover, the windrows did not also represent a food supply. Because of the association of each grid with a windrow, it was assumed that cover was held relatively constant between the grids. The differences noticed between the food and control grids on the two plantation ages, therefore should have been due to the level of the food resource and not to cover. The experimental manipulation of the food supply indicates that food, as distinct from cover, was an integral, important component of a quality habitat and important in determining Sigmodon densities. In addition to the windrows, vegetation throughout the grids also provided some cover and food. Due to succession, vegetation shifted from predominately grass and forbs to woody species. These changes undoubtedly affected both the quality and quantity of cover and food. Whereas the above discussion suggested that food became limiting with succession, field observations seemed to indicate that the vegetative cover increased with succession and was not a limiting factor. These impressions were based on an increase during the study in the size and density of woody vegetation, including briars, shrubs, and pines. The grass-forb component of the habitat decreased in response 76 to the woody growth; but when the grass and forbs died, the dead stalks still contributed to the total cover. The dead grass and forbs and the woody vegetation were not utilized for food. Sigmodon densities lend some support to these observa- tions. As previously noted, the 1972 combined food grids had higher densities than the 1973 combined food grids; yet the reverse was true for the combined control grids. This suggests that the major factor which kept the 1972 plantations from being optimal habitat was the food supply, and not cover. The high densities of Sigmodon on the food supplemented areas support the central hypothesis that food is a critical factor in the dynamics of S. hispidus during the winter and in suboptimal habitat. Mortality rates previously have been suggested to be the primary determinant of Sigmodon densities (g. g., Schnell, 1968). The present experiment indicates that changes in the natality rates rather than mortality rates may best explain the observed population fluctuations. A general relationship between reproductive activity and population density has been observed before (Goertz, 1965b; Fleharty g£_gl., 1972; Layne, 1979; Cameron, 1977). In addition, the present study provides experimental data which establish that Sigmodon natality rates are dependent on the food supply. As previously noted, granivorous rodents are thought to be limited by food more frequently than herbivorous 77 rodents. In part, the supplemental food studies of Krebs and DeLong (1965) and Cole (1977) support this statement. The present study documents that food can limit populations of S. hispidus, an omnivorous rodent. The contrasting conclusions from the various supplemental food studies may be due to different nutritional requirements and to the feeding strategies of the rodents. The design of the experi- ment may also be an important factor. Unlike most previous studies on rodent population dynamics, the present experiment was replicated which allowed the data to be analyzed statis- tically. Also, populations were simultaneously studied in habitats of different quality. As illustrated here, habitat quality can affect the results and the subsequent conclusions. Before population regulation can be fully understood, additional experimental manipulations of rodent populations are required. Also, the nutritional requirements and the availability of a quality food for rodents in the field need to be studied further. SUMMARY An experiment was designed to evaluate the following hypothesis: Food is a critical factor in the dynamics of S. hispidus populations particularly during the winter and in suboptimal habitats. Data from experimental supplementa- tion of the food supply of Sigmodon populations on six young loblolly pine plantations support this hypothesis. Data collected at the start of the experiment were used to predict that pine plantations planted in 1973 and 1972 represented optimal and suboptimal Sigmodon habitat, respectively. Population data gathered during the experiment substantiated this prediction. Densities of the 1973 control Sigmodon p0pu1ations were higher, and the breeding seasons were longer than for the 1972 control p0pu1ations. Populations from the control grids on both plantation ages exemplified demographic patterns typical for Sigmodon in South Carolina. For both plantation ages the addition of supplemental food increased population density, lengthened the breeding season, increased average litter size, increased average growth rates, and improved the physical condition of Sigmodon during the winter. Supplemental food decreased trappable longevity. Decreased survivorship on the 78 79 food-supplemented grids may have been due to an increased emigration rate in addition to low trappability of Sigmodon on the food grids. The food treatment had a greater effect on Sigmodon populations in the 1972 plantations than in the 1973 plantations. POpulation densities were higher on the 1972 food grids than on the 1973 food grids especially during early summer and winter. The effects of the supplemental food were most noticeable during the winter when the natural food supply is believed to be at a low level and when cold temperatures increase the energetic demands. In addition, growth rates for a fall cohort were higher on the 1972 food grids than on the 1973 food grids. Comparisons of the demographic data from all of the food and control p0pu1ations indicate that Sigmodon populations were limited by inadequate food resources, particularly in suboptimal habitat. The data also suggest that the food supply, as distinct from cover, was important in determining the densities in the older pine plantations. It can be concluded that food strongly influences the distribution and abundance of Sigmodon in areas undergoing rapid succession. Natality rates rather than mortality rates appeared to be the primary determinant of Sigmodon densities. Changes in density were closely associated with the rates of reproduction. Natality rates were dependent on the food supply. Reduced breeding during the winter was accompanied by and may have been caused by poor body conditions. The 80 supplemental food maintained the body condition of the animals during the winter, and a significant number of food-supplemented Sigmodon bred during the late winter. Contrasting conclusions between the present study and previous supplemental food studies may be due to different nutritional requirements and feeding strategies of the rodents. The levels of the food resource are important in the population dynamics of Sigmodon hispidus, an omnivorous rodent. APPENDIX 81 .mOpr Amy 6000 one HOV Hoppcoo on» mopmcmeoO hopuoH map was .AszsoocmuHssHm Oooampp opoz mcoHpmHsooo 30H£3 moOOOHOCH Lopes: meow on» 29H: omonpv pom mcHaampp on» mH Lopez: Ocooow 65p .OopcmHo mm: coHumpcwHo map new» onp mH goofing uprm one * O O O O O O O O H O O O O O O mImINs H O O O O H O O O O O O O O O Olmle :N sH NN NN NN HN :H s s N H H O O O mININs OH OH m O O m a m m a m N N H N OININs aa a o o m m a m a a m o a o o ananms m o m o m m o m m a m m z m H ouHuNs m a m z N N a m m N O O O O H mlmnms O H O O O O O O O O H H O O O Olmlms N: om ON om mN NN sH OH OH sH m o s O NH mINlms sH sH NH NH OH OH HH OH OH 3 m z N N O OINIms s: mN :H MH sH mH O m s w HN OH O HH HN mIHIms NH OH NH NH O m m s s s m s m mH NH *OIHIms ocsh mm: HHLQ< nope: .oom .:mO .ooo .>oz .poo .uoom .w:< mHSO oCSO mm: HHHQH moh< .0mmH mczh 0» mme HHLQ4 Eonm OHLN 3066 so o>HHm on on czocx mzpocoHHmm m56szopom mo Lopez: ESEHCHE .sH oHome 82 .mOpr Hmv 0000 One HOV Hogpcoo on» mopmcmeoO hoppmH on» Una .HmHmsoocmpHSEHm Ooaomgu mama mCOHpmHsooo QOHLB ouMOHOcH popes: 65mm on» an3 mwonpv pom wchamap on» mH hopes: Ocooom on» .OopcmHa mm: :oHpmpcmHQ map pooh one mH popes: pmpHm one * sN HO O: M3 MM O: HM ON MN OH OH HN OH HH M mIMINs M m N M H O O O H M M 3 N a O Olmle m HH O s m HH m m s m a m M M O muNuNs O O O O H H H H M M a N N H O OININN OH on sM O: MM :N MH M O s m O O a O mIHINs OH s OH NH O z N N a M N M N O O OIHINs MM sN sH MN HN HH M 3 m N H N M M M humums O m 2 O a H O O O O H H N M : Olmlms am mm mm ma ma :a o m o m a z a N o mnmums m M s m m O s s M N a N N H O Oanms OM M: MM sM O: MM :3 NN OH ON ON MN sH m M maHlMs OM :M On MM ON sH OH MH :H M O NH HH MH MH *OIHlms ocsm mm: HHLQ< 50pm: .nom .:mO .ooo .>oz .poo .paom .w5< mHSO oCSO has HHHQ< @694 .msma mess op msma aaee< Eogm OHLm Como so o>HHm on on czocx mscHQmmmom mzoszomom no pecans EOEHCHE .MH mHQmE 83 oa ma o ma ea oa m m a m m a a a a m o m a a a a a a m z a z a o a o o a m m aa o m a a a m m m M H m a a a a z a a a a a m a a a o a o m a m m s o m a a a m a m m a a m a m N M m a m a m a a a a m o m m a m z a a z m z N m s M O z N O O O a a 0 sm ma 3 m a a o em a m a a em mIMINs mlmlmw wmIHINs moon mm: HHLQ< .Lmz .oom .cmO .ooa .>oz .poo .poom .w:< mHSO oCSO has HHMQ< moHoon $694 .OsmH oQSO op msmH HHLQ< anm OHLw some so o>HHw on o» czocx mHmEEmE HHmEm Mo Lopez: ESEHCHE .mH mHQMB 89 N M N N H H H H M H m M N N H H 2 M M H O M N N N M N H O N a O H N N H H m Olmle H H H H H H m H N H H m N O H O H z N O H N H H m OININs M N H H H m H H m H o H H H H a H 2 H 2 M m z N N N M N H H M M N O H N N H O M N M H N M m M N H m OIHINs oczh Mm: HHLQ< .sz .oom .cmO .ooa .>oz .uoo .poom .w:< MHSO oCSO mm: HHMQ< moHoon mom< .A.o.ocooo .mH mHnt 85 m m m H O N N H O O a M H s OH M N a O z m m M a M M M M N H m H m H H H H N 2 H M H H E H 0 a N s s z m m H m H M N N H N N S H 0 M m m M N N H m m H m N H m m s M s O s s m s O H H H z a H M N 2 H H N H O H 0 H H N N N H N m mIMIMN mlNums mIHlms oCSO Mm: HHLQ< .pm: .260 .cmO .oom .>oz .poo .pmom .ws< MHOO oCSO MOS HHLQ< moHoon mop< .a.o.oeoeo .mH mHnt 86 m M M s M M N H H M H H m N M s H 2 H E H O M M M N N M M m ouMums M N N H H M M H N M m H m N H M H H H 2 H H O H H H M H N O m H M m O M H m OINIMs H m m M H H H N m H H H N N H M 2 n H H H O H H H O OH N H H N s s OH M H m OIHlms oQSO MN: HHHQ< .Hmz .oom .CNO .oom .>oz .uoo .poom .w:< MHSO oCSO mm: HHHQ< mmHoon onH .a.o.oeeeo .MH mHnt 87 mHmpmoHHOOH xopom mHHsszn mMeOOCOOOHQOHom ESHOOoch mMEMpHm mHHpmsHmm mzfioNMLo NCNOHHOHO mEopooz HHHmupsc m»E0poH£oo w>pmm mHOOHQMHO OO5m0H>oHn mcHanm u II COOKDZOCHQEU) a .mOHHw Amy Ooom Ocm HOV Hoppcoo on» moumcmeoO poupmH on» cam .AMHmsoocprsEHm Oooompp opoz mcoHpmHsooo QOan ouNOHOcH Hones: oEmm on» nqu omozpv pom wcHaomHu on» mH gonads Ocooom on» .OopcmHo mm: :oHumpcmHa on» Ham» on» mH Hones: pmHHO oan .H.e.ocoev .NH maoee 88 .mOHHw Hmv @000 can HOV Hogpcoo map mopmcmeoO HopooH ocp Ocm .OopcwHo mm: COHHNOQOHQ on» Ham» on» mH Lopez: 629 * MON NHN ooa MsH OOH MMH HHH sM Mo Ho OO OO MO NM M m Heooe MM MM Mo No NO MN MN MN OM MH MN ON HN NN NN O Hmooe OM NHH NM OM OM Ms MM MN OM OM sN HM MN MH M muNs Ma NH NH MH s M M M M M M M o M O ouNs MOH OOH as Ms OM Os MM MN sN HM MM MN ON OH o mums MO s: HM s: MM ON NN ON NN OH OH MH MH sH NN aosms oczw MO: HHLQ< nopmz .pmm .cmO .ooO .>oz .poo .poom .wz¢ MHSO ossh Mos HHHQ< mop< .OsMH mosh 0p msMH HHHQ< Bonn mOHHw OocHoEoo on» :o o>HHw on ow czocx mchoMmmow muomMEopoN mo Hones: EOEHQHS .ON oHomB BIBLIOGRAPHY BIBLIOGRAPHY Andrzejewski, R. 1975. Supplementary food and the winter dynamifis of bank vole p0pu1ations. Acta Theriol., 20:23- 0. Ashby, K. R., and M. A. Vincent. 1976. Individual and population energy budgets of the water vole. Acta Theriol., 21:999-512. Atkeson, T. D. 1979. Succession of small mammals on pine plantations in the Georgia piedmont. Unpublished M.S. Thesis, Univ. Georgia, Athens, 66 pp. Aydelott, D. G. 1971. Soils of the Savannah River Project, Aiken, South Carolina. State and Private Forestry, Southeastern Area. 35 pp. Baker, R. H. 1971. Nutritional strategies of myomorph rodents in North American grasslands. J. Mammal., 52:800-805. Batzli, G. O. 1979. Influence of habitat structure on a population of voles. Bull. South. Calif. Acad. Sci., 73:83-85. , and F. A. Pitelka. 1970. Influence of meadow mouse populations on California grassland. Ecology, 51:1027-1039. 1971. Condition and diet of cycling populations of the California vole, Microtus californicus. J. Mammal., 52:191-163. Bendell, J. F. 1959. Food as a control of a population of white-footed mice, Peromyscus leucopus noveboracensis (Fischer). Can. J. 2001., 37:173-209. Bigler, W. J., and J. H. Jenkins. 1975. Population characteristics of Peromyscus gossypinus and Sigmodon hispidus in tropical hammocks of south Florida. J. Mammal., 56:633-699. 89 9O Briese, L. A. 1973. Variations in elemental composition and cycling in the cotton rat, Sigmodon hispidus. Unpublished M.S. Thesis, Univ. Georgia, Athens, 71 pp- Bujalska, G. 1975. The effect of supplementary food on some parameters in an island p0pu1ation of Clethrionomys glareolus (Schreber, 1780). Bull. Acad. Pol. Sci. Ser. Sci. Biol., 23:23-28. Cameron, G. N. 1977. Experimental species removal: demographic responses by Sigmodon hispidus and Reithrodontomys fulvescens. J. Mammal., 58:988-506. Chitty, D. 1960. Population processes in the vole and their relevance to general theory. Can. J. Zool., 38:99-113. , D. Pimentel, and C. J. Krebs. 1968. Food supply of overwintered voles. J. Anim. Ecol., 37:113-120. Cole, F. R. 1977. Nutrition and population dynamics of the prairie vole Microtus ochrogaster in central Illinois. Unpublished Ph.D. Diss., Univ. Illinois, Urbana, 103 pp. Cole, L. C. 1959. The p0pu1ation consequences of life history phenomena. Q. Rev. Biol., 29:103-137. Collier, B. D., N. C. Stenseth, S. Barkley, and R. Osborn. 1975. A simulation model of energy acquisition and utilization by the brown lemming Lemmus trimucronatus at Barrow, Alaska. Oikos 26:276-299. Cottam, G., and J. T. Curtis. 1956. The use of distance measures in phytosociological sampling. Ecology, 37:951-960. DeLong, K. T. 1967. Population ecology of feral house mice. Ecology, 98:611-639. Dewsbury, D. A. 1970. Food hoarding in rice rats and cotton rats. Psychol. Rep., 26:179. Drozdz, A. 1966. Food habits and food supply of rodents in the beech forest. Acta Theriol., 11:363-389. Dunaway, P. B., and S. V. Kaye. 1961. Cotton rat mortality during severe winter. J. Mammal., 92:265-268. 91 , and S. V. Kaye. 1969. Weights of cotton rats in relation to season, breeding, and environmental radioactive contamination. Am. Midl. Nat., 71:191—155. Ferns, P. N. 1976. Diet of a Microtus agrestis population in south west Britain. Oikos, 27:506-511. Fleharty, E. D., and J. R. Choate. 1973. Bioenergetic strategies of the cotton rat, Sigmodon hispidus. J. Mammal., 59:680-692. , J. R. Choate, and M. A. Mares. 1972. Fluctuations in population density of the hispid cotton rat: factors influencing a "crash". Bull. South. Calif. Acad. Sci., 71:132-138. , M. E. Krause, and D. P. Stinnett. 1973. Body composition, energy content and lipid cycles of four species of rodents. J. Mammal., 59:926-938. , and L. E. Olson. 1969. Summer food habits of Microtus ochrogaster and Sigmodon hispidus. J. Mammal., 50:975-986. Flowerdew, J. R. 1972. The effect of supplementary food on a population of wood mice (Apodemus sylvaticus). J. Anim. Ecol., 91:553-566. Fordham, R. A. 1971. Field populations of deermice with supplemental food. Ecology, 52:138-196. Frankova, S. 1973. Effect of protein-calorie malnutrition on the development of social behavior in rats. Dev. Psychobiol., 6:33-93. Gaertner, R. 1968. Seasonal variations in the energy budgets of the harvest mouse, Reithrodontomys fulvescens and the cotton rat, Sigmodon higpidus. Unpublished Ph.D. Diss., Univ. Arkansas, Fayetteville, 199 pp. Gardner, R. H., Jr. 1975. Movement and distribution of confined and freely growing populations of cotton rats (Sigmodon hispidus). Unpublished Ph.D. Diss., North Carolina State Univ. at Raleigh, Raleigh, 71 pp. Gill, A. E. 1977. Food preferences of the California vole, Microtus californicus. J. Mammal., 58:229-233. Glass, A. R., R. Harrison, and R. S. Swerdloff. 1976. Effect of undernutrition and amino acid deficiency on the timing of puberty in rats. Pediatr. Res., 10:951-955. 92 Goertz, J. W. 1969. The influence of habitat quality upon density of cotton rat populations. Ecol. Monogr., 39:359—381. 1965a. Sex, age and weight variation in cotton rats. J. Mammal., 96:971-977. . 1965b. Reproductive variation in cotton rats. Am. Midl. Nat., 79:329-390. Golley, F. B. 1960. Energy dynamics of a food chain of an old-field community. Ecol. Monogr., 30:187-206. , J. B. Gentry, L. D. Caldwell, and L. B. Davenport, Jr. 1965. Number and variety of small mammals on the AEC Savannah River Plant. J. Mammal., 96:1-18. Grodzinski, W. 1971. Food consumption of small mammals in the Alaskan taiga forest. Ann. Zool. Fennici, 82133-136. Gunn, R. G., and J. M. Doney. 1973. The effects of nutrition and rainfall at the time of mating on the reproductive performance of ewes. J. Reprod. Fertil., Suppl., 19:253-258. Hairston, N. G., F. E. Smith, and L. B. Slobodkin. 1960. Community structure, population control, and competition. Am. Nat., 99:921-925. Hansson, L. 1971a. Small rodent food, feeding and population dynamics. Oikos, 22:183-198. . 1971b. Habitat, food and population dynamics of the field vole Microtus agrestis (L.) in south Sweden. Viltrevy, 8:267-378. 1972. Seasonal changes in physiology and nutrition of herbivorous small rodents. Aquilo Ser. Zool., 13:53-55. Hilborn, R., J. A. Redfield, and C. J. Krebs. 1976. On the reliability of enumeration for mark and recapture census of voles. Can. J. Zool., 59:1019-1029. Holisova, V. 1975. The foods eaten by rodents in reed swamps of Nesyt fishpond. Zool. Listy, 29:223-237. Hooven, E. F. 1972. Effects of clearcutting a Douglas fir stand upon small animal populations in western Oregon. Unpublished Ph.D. Diss., Oregon State Univ., Corvallis. 93 Kalela, O. 1962. On the fluctuations in the numbers of arctic and boreal small rodents as a problem of production biology. Ann. Acad. Sci. Fenn. Ser. A IV Biol., 66:1-38. Klein, D. R. 1965. Ecology of deer range in Alaska. Ecol. Monogr. 35:259-289. Kolodziej-Banach, A. 1976. Changes in fat content in an experimental p0pu1ations of the common vole. Acta Theriol., 21:255-266. Krebs, C. J. 1966. Demographic changes in fluctuating populations of Microtus californicus. Ecol. Monogr., 36:239-273. , and K. T. DeLong. 1965. A Microtus p0pu1ation with supplemental food. J. Mammal., 96:566-573. , B. L. Keller, and R. H. Tamarin. 1969. Microtus population biology: demographic changes in fluctuating populations of M. ochrogaster and M. pennsylvanicus in southern Indiana. Ecology, 50:587-607. , and J. H. Myers. 1979. Population cycles in small mammals. Advances in Ecological Research, 8:267-399. , I. Wingate, J. LeDuc, J. A. Redfield, M. Taitt, and R. Hilborn. 1976. Microtus population biology: dispersal in fluctuating p0pu1ations of M. townsendii. Can. J. Zool., 59:79-95. Lackey, J. A. 1979. Reproduction, growth and development in high-latitude and low-latitude populations of Peromyscus leucopus (Rodentia). Unpublished Ph.D. Diss., Univ. Michigan, Ann Arbor, 128 pp. Landry, S. 0., Jr. 1970. The Rodentia as omnivores. Q. Rev. Biol., 95:351-372. Layne, J. N. 1979. Ecology of small mammals in a flatwoods habitat in north-central Florida, with emphasis on the cotton rat (Sigmodon hispidus). Am. Mus. Novit., Lidicker, W. Z., Jr. 1973. Regulation of numbers in an island population of the California vole, a problem in community dynamics. Ecol. Monogr., 93:271-302. . 1975. The role of dispersal in the demography of small mammals. pp. 103-128. lg F. Golley, K. Petrusewicz and L. Ryszkowski, eds., Small Mammals: Their Productivity and Population Dynamics. Cambridge University Press, Cambridge. 99 Linduska, J. P. 1992. Winter rodent p0pu1ations in field- shocked corn. J. Wildl. Manage., 6:353-363. Lovejoy, D. A. 1975. The effect of logging on small mammal populations in New England northern hardwoods. Univ. Conn. Occas. Pap. Biol. Sci. Ser., 2:269-291. Martell, A. M., and A. Radvanyi. 1977. Changes in small mammal p0pu1ations after clearcutting of northern Ontario black spruce forest. Can. Field-Nat., 91:91-96. Millar, J. S. 1975. Tactics of energy partitioning in breeding Peromyscus. Can. J. Zool., 53:967-976. Myllymaki, A. 1969. Productivity of a free-living population of the field vole, Microtus agrestis (L.). pp. 255-265. 12 K. Petrusewicz and L. Ryszkowski, eds., Energy Flow Through Small Mammal Populations. Polish Scientific Publishers, Warsaw. Negus, N. C., and P. J. Berger. 1977. Experimental trig- gering of reproduction in a natural population of Microtus montanus. Science, 196:1230-1231. , E. Gould, and R. K. Chipman. 1961. Ecology of the rice rat, Oryzomys palustris (Harlan), on Breton Island, Gulf of Mexico, with a critique of the social stress theory. Tulane Stud. Zool., 8:99-123. Odum, E. P., C. E. Connell, and L. B. Davenport. 1962. Population energy flow of three primary consumer components of old-field ecosystems. Ecology, 93:88-96. O'Farrell, M. J., D. W. Kaufman, J. B. Gentry, and M. H. Smith. 1977. Reproductive patterns of some small mammals in South Carolina. Florida Sci., 90:76-89. Perkins, C. J. 1973. Effects of clearcutting and site preparation on the vegetation and wildlife in the flat- woods of Kemper county, Mississippi. Unpublished PhéD. Diss., Miss. State Univ., Mississippi State, 23 pp- Sadleir, R. M. F. S., K. D. Casperson, and J. Harling. 1973. Intake and requirements of energy and protein for the breeding of wild deermice, Peromyscus maniculatus. J. Reprod. Fertil., Suppl., 19:237-252. Schnell, J. H. 1968. The limiting effects of natural predation on experimental cotton rat populations. J. Wildl. Manage., 32:698-711. 95 Schultz, A. M. 1969. The nutrient-recovery hypothesis for arctic microtine cycles. II. Ecosystem variables in relation to arctic microtine cycles. pp. 57-68. lg D. Crisp, ed., Grazing in Terrestrial and Marine Environments. Blackwell Scientific, Oxford. Sloan, A. W. 1967. Estimation of body fat in young men. J. Appl. Physiol., 23:311-315. Smirin, J. M. 1970. K biologii melkih lesnych gryzunov v zimnii period. [in Russian, English summary]. Mater. Pozn. Fauny Flory SSSR, Nov. Ser., Otd. Zool. 95(LX), Fauna i ekologija Gryzunov, 9:139-150. Smith, M. H. 1971. Food as a limiting factor in the population ecology of Peromyscus polionotus (Wagner). Ann. Zool. Fennici, 8:109-121. , and R. W. Blessing. 1969. Trap response and food availability. J. Mammal., 50:368-369. , R. H. Gardner, J. B. Gentry, D. W. Kaufman, and M. H. O'Farrell. 1975. Density estimations of small mammal populations. pp. 25-53. lg F. Golley, K. Petrusewicz and L. Ryszkowski, eds., Small Mammals: Their Productivity and Population Dynamics. Cambridge University Press, Cambridge. Spurr, S. H., and B. V. Barnes. 1973. Forest Ecology. The Ronald Press Company, New York. 571 pp. Summerlin, C. T., and J. L. Wolfe. 1973. Social influences on trap response of the cotton rat, Sigmodon hispidus. Ecology, 59:1156-1159. Watts, C. H. S. 1970. Effect of supplementary food on breeding in woodland rodents. J. Mammal., 51:169-171. Wiegert, R. G. 1972. Avian versus mammalian predation on a population of cotton rats. J. Wildl. Manage., 36:1322-1327. , and J. C. Mayenschein. 1966. Distribution and trap response of a small wild population of cotton rats (Sigmodon M. hispidus). J. Mammal., 97:118-120. Zemanek, M. 1972. Food and feeding habits of rodents in a deciduous forest. Acta Theriol., 17:315-325. Zimmerman, E. G. 1965. A comparison of habitat and food of two species of Microtus. J. Mammal., 96:605-612.