Previous studies of growth in Microtus have been based almost entirely on weights. However, the weight of an individual vole may fluctuate widely over a short period, depending on pregnancy and parturition, length of time in a trap without food, availability of moisture, and other factors. In the course of my study, in 1954 and 1955, and parts of 1953 and 1956, measurements of total length, in addition to weights, were recorded for most of the voles live-trapped.

To test the accuracy of measurements, successive readings were compared in individual voles that were already of large adult size and that presumably either had stopped growing or were growing so slowly that the gain was scarcely detectable in the relatively short periods involved. For 200 such readings 33 per cent were just the same as previous records for the same animals, 24 per cent deviated by 1 mm., 22 per cent deviated by 2 mm., 15 per cent by 3 mm., 4.5 per cent by 4 mm., .5 per cent by 5 mm., 1 per cent by 6 mm., and .5 per cent by 7 mm. On the average, successive measurements varied by 1.43 mm., somewhat less than one per cent of the adult vole's total length. Occasional errors of two to four per cent were easily eliminated because for the voles used for growth records, series of measurements were available, with clearly defined trends. The occasional readings that deviated from the general trend for the individual were discarded.

Measurements were recorded along with other data in the field at the point of capture. Obtaining a reasonably accurate measurement on a live and struggling vole required patience and practice. With the thumb and forefinger of the left hand, I grasped the vole by loose skin of the nape, and simultaneously grasped the tail at a point approximately three-fourths of the distance to the tip. Then, with gentle but steady pressure, I stretched the vole out to its full length, meanwhile manipulating a millimeter ruler with the free fingers, so that the vole was pressed against it, with the nose pad at the end of the ruler.

The total length measurement is considered the best index to over-all size. The relative tail-length varies slightly between individuals, averaging approximately 22 per cent of the total length. Individuals having broken tails, or having the distal parts of their tails missing, were not included. The total length can be measured with greater accuracy than can either the head-and-body length or the tail-length separately.

Multiple captures often involved a female in oestrus and one or more males, or a female and her young, but other instances involved various combinations of sex and age groups. As many as five adults have been caught in a trap simultaneously at times when the population density was high. At such times, the meadow habitat is crossed by a maze of interconnecting surface runways and one runway may be traced continuously for 100 yards or more. Because each individual vole normally confines its activity to a small area, only a fraction of an acre, it is evident that individuals living at different places overlap in their home ranges, and also in the trailways followed in foraging. A high degree of tolerance is indicated. Where population is so sparse that the systems of surface runways comprise separate and isolated units, trapping experience has shown that one such system may harbor several or many individuals.

As direct observations on voles under natural conditions are rarely feasible, because of the animals' timidity, their utilization of concealing cover, and tendency to crepuscular habits, best evidence of social habits and underground life is based upon behavior of captive individuals. Many voles were kept in confinement for varying lengths of times, either singly or in association with others. Under such conditions there was sometimes sporadic fighting, but it was mainly defensive and serious injuries were rare. Two or more voles caught at a given spot regardless of whether they were found in the same trap simultaneously, or trapped separately within a short time, usually were completely tolerant of each other. When at rest in their container, such voles would huddle together in a corner or in a nest, if materials were provided, so that collectively they presented the minimum exposed surface. The intimacy and lack of antagonism displayed on such occasions, suggested that the voles were accustomed to living together amicably in the same nest chamber. In live-trapping, "double" captures in a single trap often involved the same two individuals. Such trap-mates were often male and female, and in many instances the female was not in breeding condition. That the voles are not monogamous in habits was demonstrated when the same female was often trapped in association with either of two males. Other trap associates taken together repeatedly often were two males, or two females. Voles that are nest mates or "neighbors" may tend to move about together in their foraging, or one confined in a trap may attract the other sufficiently to cause it to force an entrance by lifting the heavy door of a trap.

When a new vole, caught at a different location, is added to a container in which one or more are already confined, there is mutual circumspection between the original occupants and the newcomer. At first, each vole is intimidated by movements of the other, and as a result, the original occupants huddle in their established corner while the newcomer cowers in the most remote part of the container. Gradually the voles become less timid and one may approach another slowly and cautiously, to sniff at it. The vole approached may react with a show of hostility which is largely defensive. In the characteristic posture of threat for defense, the vole crouches, or rears back on its haunches, with snout elevated and incisors prominently displayed. If the warning posture is unheeded, or if the vole is made unusually aggressive by having young to defend, or for some other reason, it attacks with a sudden forward lunge, striking the adversary simultaneously with both forefeet and with the incisors. The lunge is so rapid that when I have observed it, I have been unable to discern whether the attacker bit its opponent. The attack serves to force back the other animal, throwing it off balance and intimidating it. The attacked animal may dodge nimbly to avoid the lunge, but whether or not it is actually struck, it usually retreats, avoiding or postponing further hostilities. Voles that have been kept in containers for periods of hours or days tend to be more hostile and aggressive toward a newcomer than are those newly introduced. After series of meetings resulting from the exploratory behavior of the newcomer and the curiosity or normal activity of those longer confined, hostility gradually subsides. Within a few hours a newcomer is usually accepted, and thenceforth he huddles with other members of the group when at rest, and hostility is rarely evident.

This ready acceptance on short acquaintance of strange voles into the family or social group suggests that lack of territoriality extends even to the use of the nest burrows, and that groups of voles may share the same nest, huddling together and deriving mutual benefit from the association, such as warmth in cold weather. Schmidt (1931: 113), studying this vole in Clark County, Wisconsin, noted its colonial habits. He found isolated small mounds that were riddled with burrows, and little sign in intervening areas. At one mound he trapped two adult males, one adult female, and two young; at another mound, two adult males, two adult females, and four young were trapped. My individuals that were released from live-traps were on many occasions trailed by means of a stiff wire collar with spool of thread attached, to holes that presumably were their home burrows. Data obtained in this manner indicated that ordinarily several or many individuals use the same burrow system. The histories of individual voles on the study area at the Reservation indicate shift of home base from time to time, usually for short distances within the area already included in the home range, but occasionally to new areas relatively remote from the original home range.

Severe fighting between adult prairie voles occurs at times. Occasionally, sharp squeaks accompanied by brisk rustling in the grass suggesting pursuit or conflict, are heard in their habitat. An unusually large adult male, long resident on a study area, suddenly lost weight and deteriorated in condition over a period of several days, then was found dead in a nest-box attached to a trap. Dissection revealed numerous punctures in the skin and flesh of the neck and back, probably made by the incisors of another vole. Extensive hemorrhage and swelling had occurred, and obviously these injuries were the cause of death.

Although it was not feasible to study the home life of the voles underground, clues were gained from those uncovered in runways and nests beneath large boards and strips of tarpaper, previously distributed for this purpose. Nests were constructed by the voles beneath several such pieces of tarpaper and runways appeared beneath all the pieces that were placed in habitat favorable to the voles. In summer, however, the high daytime temperatures beneath these shelters made them uninhabitable to the voles, and they were used mainly in spring. From February 15 to May 1, 1953, 14 voles were caught 19 times beneath five of the tarpaper strips, and many other voles that were seen beneath them escaped. Upon turning one of the strips I often discovered voles in close proximity. Sometimes two or more darted from the same nest. The disturbance of repeatedly raising the strips and exposing the voles' shelters soon caused them to desert the sites; consequently the information obtained by this means was limited.

There is sexual activity in every month of the year, but its incidence varies greatly from one season to another. As has been indicated by various authors, male voles reach sexual maturity later than females. It seems that ordinarily the availability of sexually active males is not a limiting factor, however. While males that are still well below average adult size produce mature spermatozoa, and are probably capable of breeding (Jameson, 1947: 145), certain large old males may sire a disproportionately large percentage of the litters produced. Observations on males in confinement indicated that sexual activity tended to be directly proportional to the size of the testes. Occasional individuals, having much enlarged scrotal testes were more readily stimulated to sexual activity and more aggressive toward females than were those in which the testes were of more nearly typical size or abdominal or were smaller than normal. The combination of factors controlling size of testes is not well understood, but males having unusually large testes were caught most often when food supply was optimum, for instance after a period of heavy precipitation when an abundant supply of new grass provided succulent and nutritious food.

In confinement sexual activity was largely inhibited and attempts to establish a laboratory colony met with failure. Sexual activity was observed mainly in recently captured males, and their interest was aroused chiefly by females that had given birth to litters within a few hours previously. Oestrus is known to follow closely after parturition. Females found in live-traps with newborn young often were brought to the laboratory for observation. An apparent instance of hostility between rival males competing for an oestrus female was observed on September 2, 1950. The female was found in a trap with four newborn young, and since the young had not yet attached to her teats, she was temporarily returned to the trap after recording, to prevent desertion of the litter. Returning twenty minutes later I found another adult vole at this trap. It would suddenly emerge from dense grass nearby, and would move over the trap or around it, with jerky, halting movements, then would dart back under cover. The female emerged from the nest box into the trap runway, and sniffed at the other, and both pressed against the intervening wire barrier. There was gnawing on the wire by one or both. A third adult vole appeared. As it moved toward the trap, all three suddenly took alarm and darted back under cover, the female hiding in the trap nest box. In a few seconds they again appeared. The two outsiders, presumably both males, were not individually recognizable, but several times one was seen to dart at the other, chasing it away momentarily. They were seldom both in sight at once.

Males confined with post-partum females usually evinced sexual interest, following them about persistently and nuzzling their genitalia. The females, however, were often unreceptive perhaps because they were disturbed by strange surroundings and by the presence of their litters, so that they usually attempted to escape, or to rebuff the male's attention. At first the female might flee, squeaking in protest at the male's pursuit. If he still continued to follow, she would turn on him, rearing back in the characteristic threatening pose, and would lunge at him, striking him sharply or driving him back. After such rebuff, males were usually intimidated or discouraged so that they temporarily or permanently abandoned their advances, and small males were more easily rebuffed than were larger individuals. On several occasions large males having enlarged testes were not readily rebuffed by females but continued to follow them. When the female turned upon him, such a male might lunge against her, throwing her off balance, and causing her to attempt to escape, and then continuing the pursuit until it ended in copulation or in more severe fighting. Although not accepted sexually, a rebuffed male might be readily accepted as a nest-mate, huddling along with the female and perhaps other individuals of both sexes. In huddling voles, the most frequently observed type of social behavior was grooming; one individual would slide its chin or muzzle through the other's fur with a stroking movement consisting of a series of rapid forward jerks and the stroking movements might continue for periods of minutes. The recipient of the grooming usually made no evident response indicative of either pleasure or displeasure. Often it seemed to be sleeping while the grooming was performed. Individuals of both sexes performed this grooming and the recipient might be of either the same sex or the opposite sex. This grooming may have some significance as a search for ectoparasites such as fleas, or mites that often infest the voles. However, after prolonged grooming by a companion, a vole's fur was of mussed and disarranged appearance. Although the grooming that occurs between voles that are resting in nests seems to have no direct significance as sexual behavior, somewhat similar actions constitute part of the mating pattern. A sexually aroused male overtaking a receptive female, slides his chin forward along her back with jerky, stroking movements. In some observed instances this behavior continued intermittently for several minutes before actual copulation. In some other instances it was almost lacking.

In female voles that are sexually quiescent, both those that have not yet attained breeding maturity, and those that have undergone regression after attainment of sexual maturity, the vaginal orifice is not evident. The canal is sealed externally by a membranous layer of epithelium. Presence of a vaginal orifice indicates that the individual is in some active stage of the breeding cycle. The appearance of the orifice varies between different females, and it changes in the same female from day to day or even from hour to hour. Presumably these changes in the vaginal orifice are cyclical and are closely correlated with oestrus, but attempts to trace them were unsuccessful largely because the normal cycle was rapidly suppressed in captive voles, which soon became sexually quiescent. Individual voles living under natural conditions were not trapped with sufficient regularity to permit tracing the details of changes in their genitalia.

In those females having the vaginal orifice most developed, the margins are turgid and slightly inflamed. The circular opening gapes 1.0 to 1.5 mm. in diameter when the tail is raised. A female may remain in this condition for two days or more. Vaginal smears at this stage often showed nucleated cells characteristic of oestrus. Subsequently the margins of the orifice become less prominent and the opening becomes smaller. The dorsal and ventral walls adhere until an opening is no longer evident unless the adjacent skin is stretched.

In pregnancy the orifice is occasionally sealed, but usually is evident. It is, however, less prominent than in oestrus, and does not gape. The margins are less turgid than in oestrus, and the opening is in the form of a transverse slit through which the purplish epithelial lining of the dorsal wall of the vagina can be seen. After parturition, placentae and bloody discharge often are in evidence in the vaginal canal. Females that have not given birth to young recently may also have bloody mucous discharge. Its significance has not been determined. In females that are undergoing sexual regression, the margins of the vaginal orifice become shrunken and pale, and the orifice becomes partly or wholly sealed.

Bodenheimer and Sulman (1946:255) concluded from their study of Microtus guentheri that in this species, as in "the cat," "the rabbit," "the ferret," and a few other mammals, ovulation is induced by copulation, and that there is no regular vaginal cycle. Hoyte (1955:412) disagreed with these conclusions for other species of Microtus, as he trapped individuals of M. oeconomus that had recently ovulated without copulation (at least no sperm were found in the genital tracts). In M. ochrogaster oestrus seems to be controlled largely by the food supply, at least the incidence of perforate females was found to fluctuate irregularly tending to follow the trend of rainfall, and, probably in more direct correlation, the amount of new grass present (see Table 1, and Martin, 1956:383-384). It therefore seems unlikely that in this species ovulation is dependent on copulation.

In females that have not yet produced young the teats are minute and well concealed in the fur, so that they are difficult to find, but in lactation they become conspicuous. In early lactation the teats are typically about 1 mm. in diameter and 2.5 mm. in length. As lactation progresses, they become thickened to nearly twice the original diameter. After lactation, as inversion occurs, they shrink to scabrous low prominences, 2 mm. to 3 mm. in diameter, surrounded by bare skin. There are three pairs of mammae, one pair pectoral and the other two abdominal. As mentioned by Jameson (1947:146), the pectoral mammae show little evidence of use in lactating prairie voles. Probably they are not used at all except in females with more than the four young in a litter accommodated by the abdominal mammae. As in various other rodents, the suckling young may cling to the female's teats and may be dragged over the ground as she moves about. When the female forages near the nest, she may drag the young with her instead of leaving them, but she can detach them instantly if she so desires. On many occasions females found in live-traps had young that were several days old clinging to their teats. In some instances young that had their eyes open may have followed the female into the trap and attached afterward.

Tables 1 and 2 and Fig. 1, based on 11,109 records representing each month over a four-year period, show the changing trends from month to month. The perforate condition recorded in Table 1 may represent any of several stages in oestrus or pregnancy, but is regarded as a crude index of rate of breeding, since voles in the anoestrus stage lack the vaginal orifice. Highest percentages of perforate females occurred in the months of February, March, April, May, and June, while by far the lowest percentages were recorded in the drought summers of 1952 and 1953. Even in mid-winter a substantial proportion of the females trapped were perforate.

Usually pregnancy can be recognized only in the last week before birth of the litter, when the female's abdomen is noticeably distended by the enlarged fetuses. Palpating to detect embryos was not attempted because of the danger of injuring them or the female. Because gestation is of approximately three weeks duration, the figures in Table 2 represent roughly perhaps one-third, or a little less, of the adult females actually pregnant. At most times of year a substantial proportion of adult females (sometimes nearly all) are pregnant. Only in the winter (including March in 1951) were samples taken in which no recognizably pregnant females were found. Incidence of pregnancy was notably high in July, August, September, and October of 1950, May, 1951, May, 1952, and April and May, 1953. A high rate of breeding was not necessarily followed by an increase in the population. A relatively low rate of breeding was adequate to maintain the population level, provided that environmental factors remained favorable. Fig. 1 shows the average catch per day (with approximately 100 live-traps) over the four-year period, 1950 through 1953. The young (including all those weighing 30 grams or less, and corresponding roughly with the part of the population less than two months old) are shown separately. It is noteworthy that throughout the entire period the ratio of young to adults tended to be fairly stable—usually fluctuating between ten and thirty per cent of the total catch. Ratios of young to adults were notably high in March and May, 1950; April, June and July, 1952; and April, May and June, 1953. Ratios of young were notably low in June and December, 1950; January, February, March, and June through October, 1951; January, February, and March, 1952; and November, 1953.

In Fig. 1 the catch per day of voles, varying from month to month, reflects chiefly the changing population density. However, other factors also have important effects on the catch. For example, bait acceptance is better in the winter when natural foods, especially greens, are scarce, with the result that a higher catch can be made with the same population density. Interference with the trap line by other animals also affected the catch of voles. In warm weather the traps were checked in both morning and evening, and the catch was correspondingly greater than it was in cool weather when the traps were checked only once daily. The ratios obtained of young to adult voles cannot be accepted at face value as the true ratios in the population, either. For the first several days of each trapping period, the voles caught were mostly adults previously marked and, presumably, conditioned to the grain bait. Later, young voles not previously recorded, came to the traps in increasing numbers. The young, being at first not conditioned to the bait, and also having relatively small home ranges, would generally be less well represented in the catch than would the adults.

A female live-trapped on July 23, 1951, appeared to be in breeding condition. When trapped two days later, she had a copulatory plug, and 21 days after this she was found with a newborn litter in a trap. A female thought to have given birth to a litter between successive captures on July 20, and July 21, 1951 (on the basis of appearance of genitalia, and reduction in weight from 53 to 46 grams), appeared to have just completed parturition when she was examined on August 10. A female that gave birth to a litter in confinement on May 18, 1954, bred and was released the same day. She was recorded as pregnant in the first week of June, but on June 7 was no longer pregnant. If this pregnancy terminated normally, a gestation of 20 days or less is indicated.

Greenwald (1956:221) suggested that in M. californicus, oestrus might occur in the period of lactation, because he found recently formed corpora lutea in lactating females. In the course of my field work on M. ochrogaster, I obtained precise or approximate dates of successive litters born at intervals of somewhat more than 21 days apart. In different females, intervals of 23, 23, 24, 26, and approximately 27 (between 26 and 28) days were recorded between successive litters. In four other females intervals between litters were known only approximately because one of two records was based on a capture in late pregnancy judged to be within two or three days of parturition. For these females, intervals of 23, 24, 24, and 26 days were recorded. From the trend of these records, it seems that females often became pregnant within a few days after birth of a litter. Pregnancy from post-partum oestrus would seem to be less frequent than pregnancies beginning a few days after birth of the previous litter, and within the period of lactation.

Mean numbers of young per litter were found to vary from year to year and from month to month, as shown by the following lists: 1950, 3.0 (13 litters); 1951, 3.5 (23 litters); 1952, 3.5 (11 litters); 1953, 3.4 (5 litters); 1954, 3.4 (15 litters); 1955, 4.1 (7 litters); 1956, 3.8 (5 litters); January 2.0 (1 litter); February 3.5 (4 litters); March 4.5 (4 litters); April 3.9 (12 litters); May 3.3 (25 litters); June 3.0 (9 litters); July 2.7 (4 litters); August 2.9 (7 litters); September 2.8 (6 litters); October 3.4 (7 litters); November 5.0 (2 litters); December 4.0 (1 litter).

These differences can be logically explained on the basis of changes in the average age of the breeding females in the population. On the average, with greater length, weight and age, females produced progressively larger litters, although individuals did not necessarily conform to this general trend. For 24 females recorded in 1954-1956 and measured within a few days of birth of their litters, average length was correlated with number of young as follows: 6 young, 163.5 mm.; 5 young, 158.0 mm.; 4 young, 157.7 mm.; 3 young, 154.6 mm.; 2 young, 160.5 mm.

For 48 other females, recorded in 1950-1953, that were not measured, but that were mostly assignable to broad age groups on the basis of their individual histories in the trapping records, the following well defined trend was demonstrated.

In four newborn young, total lengths, in mm., were 47, 45, 45, and 42. From the length-weight relationships shown in Fig. 2, it seems that a length of approximately 47 mm. is typical of newborn young of average weight. Martin (1956:388) found a mean weight of 2.8 ± 0.36 grams in sixteen newborn prairie voles from the Reservation. For a series of 67 other newborn voles representing 27 different litters in seven different years, I found an average of 2.9 ± .05 grams. Young ranged in weight from 3.8 to 2.0 grams. Weights of the newborn voles could not be correlated with season, size, age of females, or food conditions. However, a distinct trend toward larger size in those litters that contained fewer young was evident, as shown in Table 4.

Newborn voles were obtained when females that were caught in live-traps produced their litters before they were found and released. In some instances, females caught while in late pregnancy were retained in the laboratory for a day or more until parturition occurred. Many of the newborn voles were marked by toe-clipping, according to the same system used for adults. Early growth was measured in some instances by keeping the female with her litter in confinement, measuring and weighing the young at intervals. In most instances, the female was released at the point of capture (presumably near her nest burrow) with the young clinging to her teats. For the young so released, the incidence of recovery was remarkably low, seeming to indicate that they were subject to decimating losses. Perhaps such losses are normal, at least on the study area where voles are live-trapped regularly. Holding of adults and partly grown young in live-traps ordinarily has no harmful effects on them, but the resultant separation of females from newly born litters may often result in death of the young either from hunger and exposure, or from attack by other voles and natural enemies.

During the first ten days the increase in length from an original 47 mm. is from three to four mm. per day. Figs. 2, 5, and 8 show length and weights of voles whose ages in days were definitely known because they were born in the laboratory, or in a live-trap after the female was caught there. Young voles marked at birth and released with the female were rarely recovered in the period of suckling, as they ordinarily remain in the nest burrow when the female ventures out to forage. Litters retained in the laboratory therefore have provided most of the records of growth in suckling young. Growth varied greatly between litters. It was not clearly correlated with size of female, size of young at birth, or number of young in litter, but probably was influenced by attentiveness of the female, her adjustment to captivity, and her productivity of milk. Within each litter there were usually persistent differences in development, but these were minor (except for those of occasional runts) compared with the differences between litters. In several litters of five young, one was usually smaller than the others at birth and therefore could not compete successfully with its litter mates, so that it never gained possession of a teat other than one of the pectoral pair, and always succumbed within a few days, after failing to gain weight as its litter mates did. The relatively few voles marked at birth and recovered after developing under natural conditions, did not deviate from the trend of those in confinement.