m HARVARD UNIVERSITY Library of the Museum of Comparative Zoology Number 1 V^, 6« August 10, 1936 ■VI Editor's Note; The publication difficulties which have lately beset all educational and scientific institutions are mani- fest. Coincident with reduced publication allowances and increased printing costs there has been a growth in research activity. The manuscripts submitted have far exceeded the printing budgets, necessitating the rejection or indefinite postponement of much worthwhile material. The result has been a loss to science and discouragement to the individual performing the work. The San Diego Society of Natural History, as a partial remedy for this situation in its own case, begins herewith the publication of a third series of scientific papers to be known as the Occasional Papers. In this advan- tage will be taken of the newer and cheaper lithostatic pro- cess of facsimile reproduction. While not having the legi- bility and perfection of form of standard printing it is be- lieved that it will be entirely satisfactory for the charac- ter of papers contemplated for this series. These will be reports primarily of interest to specialists, rather than the monographic type appearing in the Society's Memoirs or the descriptive and taxonomic papers contained in its Transact- ions. The two latter series will be continued as in the past. A STATISTICAL STUDY OF THE RATTLESNAKES By LAURENCE M. KLAUBER Curator of Reptiles and Amphibians, San Diego Society of Natural History I - INTRODUCTION Snake dens, or winter rendezvous, where numbers of snakes congregate in the autumn prior to hibernation, or emerge in the spring, are well authenticated. Amongst the many incredible tales of the serpent-tribe this, at least, has been attested by many reliable witnesses. A number of persons, hearing of my interest in snakes, and particularly rattlesnakes, have written me of experiences with these dens. Some of these stories were no doubt true, others had grown considerably with the tel- ling; but all lacked concrete verification. Finally, one such letter was received October 17, 1950, from my good friend C. B. Perkins, then of Denver. In this he recounted his experiences in raiding a gathering place of the prairie rattlesnake in a prairie dog town east of Platteville, Colorado. This town he had been watching during the preceding summer and fall, and had finally been rewarded by the congregating of the snakes, of which he had succeeded in capturing a considerable number. But in its conclusion Mr. Perkins' letter differed from those previous- ly received, for he wrote, "I am sending you today, by express, one hundred and fifty- three of these snakes." Here, at last, was what I had hoped to secure — a group of specimens from a single locality, collected within the space of a few days. This shipment was the beginning of a series of raids on this and several adjacent prairie dog towns used by the rat- tiers as winter quarters, and a traffic in snakes began which did not terminate until Mr. Perkins' return to California in the spring of 1932. During this period I received, through his kindness, no less than 804 rattlers, all of which were preserved for study; and dimensional data on 57 additional specimens, which he had retained xuader observation alive or had sent to various scientific institutions, were made avail- able to me. Thus, there is at hand a fairly complete statis- tical record of 861 specimens of the prairie rattlesnake, Crotalus confluentus confluentus, collected within a compact area and representing a complete cross-section of one rattler community. Mr. Perkins' generosity has involved an obligation. A worker in any field of natural history invariably chafes under the handicap of lack of material; always there are ques- tions just beyond solution because specimens from the right place, the right season, or sufficient in quantity are not at hand. But here was a series which ought to answer at least a few of the problems involving the rattlers of east-central Colorado; certainly one could not plead lack of material. They seemed to afford an opportunity to determine which of the obvious variations of the rattlers, in form and lepidosis, might be attributable to individual or sexual differences, and which to differences of species or habitat. It was hoped that these Colorado specimens (the Platteville series as they -^ are called hereafter) would constitute a datum, or basis of '^,\^ comparison, for each differential character, which might be •<^'-f employed in evaluating the variations found in other series of the same subspecies, and thence the investigation might be extended into the other subspecies and species of the genus. (2) This has now been done, at least for some of the characters used in classification, and for certain morpho- logical elements. While fiindamentally it may be said that Mr. Perkins' Piattevilie series of confluentus constitutes the basis of these notes, corroborative data were secured from others, amongst which may be mentioned especially a series of confluentus from central South Dakota, of which 673 are in my own collection; these were secured through the courtesy of Mr. A. M. Jackley, who collected them in the field, and Mr. Howard K. Gloyd. There is likewise a series of 1S9 juvenile C. atrox from San Patricio, southern Texas, procured from Mr. H. J. Yoder; a series of 294 C. lucasensis brought in by Mr. Fred Lewis, from Cape San Lu- cas, Lower California; and a group of 24.4 C. c. oreganus from near Pateros, Washington, collected by P.J.Martin and others. I am also greatly indebted to many other indivi- duals and institutions who have contributed specimens by gift or loan, so that I have been enabled to carry the in- vestigation into all the rattlesnake speciesj to others I am obligated for information and data of various kinds. To all of these I hope to make acknowledgment elsewhere. Al- together scale counts and measurements have been available on more than 8000 rattlesnakes. Some 10,000 specimens have been seen alive, including all of the recognized forms except unlcolor. polystictus, ste.inegerl. and ravus. This is not a monograph of the rattlesnakes, nor an attempt particularly to unravel the problem of the gene- tic relationships of the species. Rather it is a tracing of the trends of a number of different variables through the group, for a survey of the dispersion of characteristics of form and lepidosis constitutes the essential object of the investigation. If, thereby, relationships are indicated or determined, incidental steps toward the solution of the taxonomic problems will resultj but primarily the purpose has been to ascertain the extent and equations of the variations; how they cluster about the mean; and whether they are indivi- dual, sexual, or specific. It is hoped, also, that through this study of the rattlers, some indication may be given of the general value of certain morphological characters as dif- ferential indicators amongst the snakes, particularly those which are ordinarily employed in ophidian classification. I am duly aware that this is a presiimptuous program upon the part of one who has had no formal training in either statis- tics or herpetology. Always the searcher hopes to find, perhaps hidden in some inconspicuous character, a definite and invariable key, distinguishing or individualizing one form from all its relatives, as if nature had written here the technical name of that form. But usually what is available is only the cumulative evidence of many characters, few of which coincide at points of divergence. Some of these may be translated into the statistics of average and correlation, and thereby serve, by weight of evidence rather than by rigid limits, to differentiate the forms. These cumulative weights are diffi- cult to determine, and thus individual classification or group segregation becomes difficult. There is confusion between those variations produced by directional trends in form and others which are the result of individual or sexual differen- ces. There are still other divergences, which, while defin- ite as trends, are too unstable or too vague in value to war- rant even the smallest definable distinction, that is, a divi- sion into subspecies or geographical races. (3) Differential characters may be of a type not des- cribable as positive or negative, present or absent, nor in numerical or statistical form. An investigator of long ex- perience with a certain animal, notes automatically these dif- ferences and their cumulative effectiveness, or weight; but they are difficult to explain or evaluate for the use of other workers with less acquaintance with that particular genus or family. Thus, the large series becomes important in order that the weight and dispersion of the many characters, which in summation, constitute the basis of differentiation, can be followed through from the Individual to a family or brood, thence to a localized group, then to collections of wider geo- graphical range and varying habitat conditions, and thus to the classifications of subspecies, species, and genus. Mean- while the parallel differences due to sex and age must be no- ted and correlated with the others. By such studies the val- ues of the several variants as key or reference characters may be determined. This involves also a study of' the relative stability of characters. Large series make it possible to determine av- erages of the numerical data with relatively small probable errors of the mean, that is, the errors inherent in random sampling. Having determined these, as well as the disper- sions of the variates, we may next check the significance of the differences (for various characters) between two groups, such as members of the same subspecies from different geograph- ical locations, or ecological niches, to see how far diver- gence has gone and which characters show the highest degree of plasticity or mobility, and which, conversely, of stability. These data in turn should aid in the evaluation of subspecific and specific differences, thus indicating, by composite weight, relationships and the validity of taxonomic separation; for it would seem that differences in stable characters should be giv- en more weight in judging species differences, than deviations in characters shown to be unstable within a species. In discussing the interrelations and variations of a genus, either the characters, or the species may be treated con- secutively. In many instances both methods are desirable, since only by this means are the relative values of the dif- ferences clearly evident; this, however, involves much duplica- tion. In the present notes the character, rather than the species, is taken as the unit, the investigation being carried forward under such subjects as length, weight, head and tail proportions, fangs, venom, rJittles, hemipenes, scale counts, etc. In each case some attempt is made to tabulate the var- iations by species; but the corollary presentation, that is, a listing of characters under each species with a complete species description, is, for the present, omitted. For purposes of clarity a list of the species and subspecies considered valid is presented below: 1. Crotalus durissus durissus Linne', 1758. Central American Rattlesnake 2. Crotaius durissus terriflcus (Laurenti), 1768. South American Rattlesnake 3- Crotalus unicolor de Jeude, 1887. Aruba Island Rattlesnake {A) 4. Crotalus baslllscus (Cope), 1864. Mexican West-Coast Rattlesnake 5. Crotalus enyo (Cope), 1861. Lower California Rattlesnake 6. Crotalus molossus molossus Baird and Girard, 1853. Northern Black-tailed Rattlesnake 7. Crotalus molossus nigrescens Gloyd, 19S6. Southern Black-tailed Rattlesnake 8. Crotalus adamanteus Beauvois, 1799 Eastern Diamond Rattlesnake 9. Crotalus atrox Baird and Girard, 1853. Western Diamond Rattlesnake 10. Crotalus tortugensis Van Denburgh and Slevin, 19E1. Tortuga Island Diamond Rattlesnake 11. Crotalus lucasensis Van Denburgh, 19£0. San Lucan Diamond Rattlesnake 1£. Crotalus ruber Cope, 1892. Red Diamond Rattlesnake IS. Crotalus exsul Garman, 1883. Cedros Island Diamond Rattlesnake 14. Crotalus scutulatus (Kennicott) , 1861 Mohave Rattlesnake 15. Crotalus confluentus confluentus Say, 1823. Prairie Rattlesnake 16. Crotalus confluentus nxmtius Klauber, 1935. Arizona Prairie Rattlesnake 17. Crotalus confluentus abyssus Klauber, 1930. Grand Canyon Rattlesnake 18. Crotalus confluentus lutosus Klauber, 1930. Great Basin Rattlesnake 19. Crotalus confluentus concolor Woodbury, 1929. Midget Faded Rattiesn&ke 20. Crotalus confluentus oreganus Holbrook, 1840. Pacific Rattxesnake 21. Crotalus mltcheilil mitchellii (Cope), 1861 San Lucan Speckled Rattlesnake 22. Crotalus mitchellii pyrrhus (Cope), 1866. Southwestern Spe"ckled Rattxesnake 23. Crotalus mitchellii stephensi Klauber, 1930 Panamint Rattlesnake 24. Crotalus tigris Kennicott, 1859. Tiger Rattlesnake (5) 25. Crotaius cerastes Halioweli, 1854. Horned Rattlesnake j Sidewinder 26. Crotaius polystictus (Cope) 1865. Mexican Lance-headed Rattlesnake 27. Crotaius horridus horridus Linne, 1758. Timber Rattlesnake 28. Crotaius horridus atricaudatus Latreille, 1802. Cane-brake Rattlesnake 29. Crotaius lepidus lepidus (Kennicott), 1861. Eastern Rock Rattlesnake 50. Crotaius lepidus klauberi Gloyd, 1936. Green Rock Rattlesnake 31. Crotaius triseriatus trlseriatus Wagler, 1830. Mexican Spotted Rattlesnake 32. Crotaius triseriatus prlcei Van Denburgh, 1895. Arizona Spotted Rattlesnake 33. Crotaius stejnegeri Dunn, 1919. Long-tailed Rattlesnake 34. Crotaius willardi Meek, 1905. Ridge-nosed Rattlesnake 35. Sistrurus ravus (Cope), 1865. Mexican Groiind Rattlesnake 36. Sistrurus miliarius miliarius (Linne^ ), 1766. Carolina Groiind Rattlesnake 57. Sistrurus miliarius barbouri Gloyd, 1935. Southeastern Ground Rattlesnake 38. Sistrurus miliarius streckeri Gloyd, 1935. Western Ground Rattlesnake 39. Sistrurus catenatus catenatus (Raf inesque) , 1818. Eastern Massasaxiga 40. Sistrurus catenatus tergeminus (Say), 1823. Western Massasauga It is to be expected that some of the above may be segregated into additional subspecies as further studies by workers in this field are carried forward- Also changes in names may be found to be necessary, as is indicated by studies in nomenclature which the writer now has under way. In the presentation of some of the statistical data I have made some segregations between geographical groups of the same subspecies where there is a possible (or incipient) race not yet accorded recognition, as, for instance. Coronados Islands oreganus. Also, certain of the large serieo already mentioned are occasionally shown separately in order that the dispersion of characters in territorially restricted groups may be presented. (6) The order of publication of the several subjects discussed in this series is without logic or significance. Each is more or less complete in itself and publication will depend on the completion of the necessary data and the pub- lication program of the San Diego Society of Natural History. In the course of these Investigations, new facts have already come to light on certain taxonomic problems of the rattlers. These have resulted in the description of new subspecies, or the reviving of forms previously thrown into synonymy. As these papers have appeared in other publica- tion series the references are given here. 1930 New and Renamed Subspecies of Crotalus confluentus Say, with remarks on Related Species. Trans. S.D. Soc. Nat, Hist., Vol.6, No. 3, pp. 95-144, plates 9-12, map. Describes C.c.lutosus. C.c.stephensl (now considered Cm. Stephens!) , C.c. decolor 1^ synonym for C.c.concolor Woodbury), C.£.abxssus. Discusses C.scutulatus. C.c. confluentus, C.c.mitchellii, and C.c.oreganus. 1930 Differential Characteristics of Southwestern Rattle- snakes Allied to Crotalus atrox. Bull. Zool. Soc. S.D., No. 6, pp. 1-72, plates 1-6, maps 1-5. Discusses atrox, tortugensis, lucasensis, ruber, exsul, and scutulatus. 1931 Crotalus tigris and Crotalus enyo. Two Little Known Rattlesnakes of the Southwest. Trtns. S.D. Soc. Nat. Hist., Vol.6, No. 24, pp. 353-370, plate 23, map. Dis- cusses these and related species. 1935 A New Subspecies of Crotalus confluentus, the Prairie Rattlesnake. Trans. S. D. Soc .Nat. Hist. , Vol.8, No. 13, pp. 75-90, plate 3, map. Describes C.c.nuntius. 1936 Crotalus mitchellll, Tne Speckled Rattlesnake. Trans. S. D. Soc .Nat. Hist. , Vol.3, No. 19, pp. 149-134, plates 19-20, figs. 1-3, map. Discusses C.m.mitchellli. C.m. pyrrhas, and Cm. stephensl . (7) II - SEX RATIO IN RATTLESNAKE POPULATIONS It has been my experience in examining preserved specimens of rattlesnakes contained in museum collections, to find a rather marked predominance of males. This condition applies to practically all species; and where sufficient speci- mens are at hand from different areas to afford a representative figure, the same situation will be found to exist in each sep- arate area. Table 1 shows the sex ratios by species of 82^0 rattlers which were investigated. The 233 specimens of which the sex was not recorded represent heads, skins, em- bryos, and similar specimens of which the sex was indeter- minate, together with some of which the sex was not noted. Subspecies are segregated only in the cases of confluentus and mitchellil; of the others not enough specimens £re avail- able to warrant separation for this particular study. The table shows a n\imerical superiority of the males in all but one species. Where the available number of a species is sufficiently large to afford representative samples, say 100 or more, it will be observed that the sex ratio varies from 1.1 to 1.6 in favor of the males, the aver- age being 1.3. Only three forms fall beyond this rangej horridus is conspicuously low, and nuntius and pyrrhus un- usually high. So consistent are these results there can be no doubt that the males collected outnumber the females. Various explanations may be suggested to account for this unbalance of the sexes. There may be more males born than females, or there may be a higher mortality amongst the females. Again, since these figures merely indicate accumulations in collections, they may not be representative of the true situation in the wild, but rather show that fe- males are less often captured than males. This, in fact, would be expected on the assumption that the males are less careful of concealment when searching for females in the spring mating season; or the females may be more secretive when heavy with eggs, and thus less often captured.* But a somewhat opposing theory would suggest that the females, if less active, would be more subject to capture and destruct- ion. We may check the original sex ratio by an investi- gation of broods at birth. Records are avbiisbie of 56 broods, 15 subspecies being represented. These broods con- tained 463 young snakes, or an averag,e of 8.3 per brood, the minimum being £ and the maximum 18. Some had been born, others were unborn embryos. The sex can be determined in 453 yo\ing and it is found that there are £13 males and £40 females, or a male ratio of 0.89. Even allowing for errors in sexing these juvenile specimens, which, with the exception of the embryos (these have the male organs extruded), are more difficult to sex than adults, the result is conspicuous- ly different from the general average shown in Table 1. » Dr. F. N. Blanchard informs me that the latter is not true in certain genera of snakes ^^hich he has investigi ted, (B) Table 1 SEX DISTRIBUTION species or Subspecies Males Females Uncertain Total Ratio ll/F Durissus 59 54 7 120 1.19 Unicolor 1 1 — Basiliscus 48 44 2 94 1.19 Enyo 39 22 61 1.77 Molossus 159 120 8 287 1.33 Adamanteus 26 16 12 54 1.62 Atrox 399 284 38 721 1.40 Tortugensis 21 7 28 3.00 Lucasensis 198 149 2 349 1.33 Ruber 154 118 16 288 1.31 Exsul 17 4 21 4.25 Scutulatus 234 143 20 397 1.64 Confluentus 1105 964 20 2089 1.15 Nuntius 122 63 3 188 1.94 Abyssus 18 12 2 32 1.50 Lutosus 229 157 7 393 1.46 Concolor 13 9 22 1.45 Oreganus 795 594 28 1417 1.34 Mitchellii 57 29 6 92 1.97 Pyrrhus 133 60 16 209 2.22 Stephens! 42 23 4 69 1.83 Tigris 26 15 6 47 1.73 Cerastes 180 140 6 326 1.29 Polystictus 9 8 1 18 1.12 Horridus 66 106 1 173 .62 Lepidus 90 71 2 163 1.27 Triseriatas 101 80 11 192 1.26 Stejnegeri 3 3 — Wiliardi 15 13 A 32 1.15 Ravus 10 1 3 U 10.00 Miltarius 116 104 7 227 1.11 3atenatus 57 55 1 . 113 I.OjL Total 4542 3465 233 8240 1.31 In addition to these broods, we have available sev- eral series of young of the year. Some of these were collect- ed in raids on dens and are typical of yoimg snakes going into their first hibernation or emerging therefrom. Others are groups of young Just after birth but which could not be included amongst the single broods, since the mothers had not been segre- gated and the young therefore were all mixed together. Also there is a series of oreganus from San Diego County, California, where for some years I have been collecting Juveniles for study. Included in this lot are all available specimens under 400 mm. in length, mostly collected in the autximn, but a few in the spring* All are young of the year. It may be assvimed that they are in much the same category as a brood; there has not yet been an opportianity for any destructive or other selective forces or habits to cause an artificial sex ratio. In the following summary the broods previously men- tioned are also included: Sinaloa series of baslliscus Mendocino series of oreganus Smith-Dunkie series of nigrescens San Patricio series of atrox Jackley series of confluentus Platteviile series of confluentus San Diego series of oreganus Total, - - - - 56 Assorted broods Hales Females Ratio M/F 10 14 0.71 12 15 0.80 44 39 1.13 77 62 1.24 80 72 1.11 125 106 1.16 102 107 0.95 448 415 1.08 213 240 0.85 661 655 1.01 From these figures it will be observed that there is a slight predominance of males amongst the young of the year. This almost balances the excess females in the broods so that when we combine the young from all sources, totaling 1316 snakes, there is almost exact equality, the male excess being only one per cent. Having in mind the variations to be expected in sampling of this kind, we feel there is every reason to believe that the sexes are born in equal nvimbers. Coming now to the adults, we may next determine whether the discrepancy between the sexes is a real one or whether it is due to those circumstances of life function and habit which lead to the more ready capture of the males. If the latter be the case, the sexes should be found in more nearly equal numbers after the time of year of active sexual fvtnction. Thus the season of retirement or hibernation should offer a clue, and we are led to consider the Platteviile series as they enter and is- sue from hibernation. For all specimens exceeding 400 mm. in length we have the following situation: (10) Season Males Females Ratio M/F Spring 104 98 1.06 Autumn £28 195 1.17 Grand Total 332 293 1.13 We see that, although there is a rather definite predominance of males, it is not nearly so pronounced as in the case of the specimens listed in Table 1, where the sex ratio, exclusive of the young already considered, is 1.40 in favor of the males. In the Jackley series there were 552 adolescents and adults J of these there were 292 males and 260 females, a sex ratio of 1.12. These also represent autumn and spring collections about hibernating retreats. It will be observed that this agrees closely with the Platteville figure. It may be suggested that sex characteristics might still be effective in causing a selection of snakes about hiber- nating dens, particularly in the spring, for the more active and ardent males might be more ready to come out daily for the first warmth of the season. This may partly account for the statistics of certain smaller spring raids on dens as indicat- ed in the following schedule: 24 C.c.lutosus (15M, 9F) taken by E. Raymond Hall May 22, 1929, 5^ Mi. SW.of Osceola, Nevada. 24 C . c . lutosus -oreganus intergrades, (16M, 8F) taken by A. L. Hagar, 16 Mi. WNW. of Coxoncii, Idaho, May 21, 1932. 23 C.c.lutosus (IIM, IgF) taken by C.B.Perkins 10 Mi. m. of Abraham, Utah, May 12, 1929. (6 on a succeeding date) . 225 C.c. oreganus (131M, 94F) taken by P. J. Martin and others, 6 Mi. N. of Pateros, Washington in May. Here we have a total of 298 snakes with 173 males and 125 females, a ratio of 1.38, which is as high as the general average of Table 1. However, it is important to observe that in two instances (including the last) there was a conscious sel- ection of the larger specimens, thus procuring more males. Del- iberate selection of this kind must be guarded against in such studies or the results will be distorted. Also the raids should be made on several consecutive days (as was the case in the Plat- teville and Jackley series) or a difference in activity may af- fect the relative numbers of each sex secured. Certain figures are at hand which indicate that there is a real seasonal difference in the frequency of capture of the sexes. In San Diego County records are available of 123 adult specimens of oreganus . We find that the sex ratio of the spring specimens (31 males, 23 females) is 1.35, while for the summer and autumn specimens (June to December) it is 2.29, there being 48 males and only 21 females. (11) Similarly, in a group of 122 adult C. scutulatus from Arizona, the spring ratio (18 males, 16 females) is 1.12, while the summer-autumn ratio (61 males, 27 females) is 2.26. In a collection of 99 adult C. atrox from Arizona the sjring ratio (15 males, 14 females) is 1.07 and the summer- autumn ratio (44 males, 26 females) is 1.69. There is in these figures a rather definite indica- tion that in the spring mating season there is a slight pre- dominance of males; later in the summer, as the females become increasingly heavy with young, they seek seclusion and are not so often collected, hence at this time the males more strongly predominate. Whether there are any species differences in sex ratio cannot be conclusively determined from the figures of Table 1, since the species which differ greatly from the mode are not represented in sufficient numbers to prove the point. It is possibly significant that all three of the mitchellii subspecies and their near relative, tigris, have abnormally high male ratios; it will be interesting to see if this holds for larger series. The low figure for horridus is somewhat more doubtful. It may be noted that the only two exclusively Island species (except unicolor, of which only one specimen was avail- able), that is, tortugensls and exsul. have very high ratios of males to females, 3.00 and 4.25 respectively. Yet Coronados Islands oreganus, the only other form of which more than 10 island specimens are available, has an unusually low ratio, there being 11 males and 20 females, a ratio of 0.55. Evident- ly these rather small island series have been affected by the season of collection; at least we are not yet in a position to conclude that island sex ratios differ from those existing in mainland races. Returning to a discussion of the 56 broods it may be stated that they vary in size from 2 to 21. A test for the significance of correlation between size of brood and percent- age of males indicates that there is none--at least it is not shown in this relatively small number of broods. It is true that the only three broods of a single sex (with 2, 3, and 4 individuals respectively) have all females, but I do not con- sider this more than a result of chances in sampling. To prove anything on this point would require a large number of broods of a single species, since it is known that the size of the brood varies both with the size of the individual moth- er and the average size of the species, smaller races having smaller broods. An attempt was made to check the sex distribution against a point binominal curve of equal sex distribution by allocating broods regardless of size to the percentage distri- bution of an 8-vinit litter, this being the average size. But the method is of doubtful validity and the number of broods too small and too varied in size really to prove (as the re- sults seem to indicate) that broods run somewhat higher in unbalanced distributions than is to be expected on the theory that the sex of each individual is the result of pure and in- dependent chance. The following table contains the distribu- tion: (12) M/F 8/0 7/i 6/2 5/5 4/4 5/5 £/6 1/7 0/8 Theoretical frequency 0.£ 1.8 G.l 12. £ 15.3 12.2 6.1 1.8 0.2 Actual fre^iuency 0.0 4.0 6.0 Ij.O 12.5 7.0 8.0 5.5 5.0 If we concede a numerical inferiority of the male") in the ratio of .47 to .53 (as indicated by the broods but not by the other young of the year) we have a somewhat closer agreement between the theoretical and actual distributions: M/F 8/0 7/1 6/2 5/3 4/4 5/5 2/6 1/7 0/8 Theoretical frequency 0.1 1.2 4.8 10.8 15.2 13.7 7.4 2.5 0.3 Actual frequency 0.0 4.0 6.0 10.0 12.5 7.0 8.0 5.5 3.0 But the unbalanced broods are stilL high at the ex- pense of the more evenly distributed; that is, the 7/1, 6/2, 2/6, and 1/7 broods are all well above the ex^vected frequency. Taking all of these data into consideration I am of the opinion that the following conclusions may be drawn with respect to the sex ratio amongst the rattlesnakes: 1. The sexes of rattlesnakes are born in approximately equal numbers. 2. There is a slightly higher mortality amongst the females which leads to an excess amongst adult males of about 10 per cent above the females. 3. Owing to the greater activity of the males, particularly in the summer and early autumn when the females are heavy with young, col- lectors capture more males than females, so that the adult males in collections average about 40 per cent in excess of females. (13) III. BIRTH RATE Age Classes - Plattevllle Series; The Plattevllle, Colorado, series of Crotalus confluen- tus confluentus offers an opportunity to make approximations of the birth rate of this species of rattlesnake, based on the num- bers of eggs found in the females. Also, as all the snakes were collected when entering or Issuing from hibernation, and without conscious selection as to size, it may be assumed that we have an average sample of all population elements. We can therefore de- termine the proportion of juveniles remaining at the first hiber- nation. A statistical presentation of the data available is given in Table 2. These data represent a study of 84I out of 861 snakes comprising the Plattevllle series; egg records were not available on twenty. These 84.I specimens may be segregated as follows: Males Females Total Juveniles (O - 399 mm.) 123 Adolescents (4OO - 579 mm.) ... 4I Adults (580 mm. or over) .... 285 Total 4-49 The reason for dividing the adolescents from the adults at 580 mm. will appear hereafter. Since these specimens were collected at the beginning or end of the hibernating season, the series may be considered as representing a single period in the life cycle, this being one during which the adult females should be carrying eggs. Life Curve Summary; Before presenting the birth rate calculations it will be advisable to consider the probable life curve of the species. As our specimens represent only a single period we must draw our conclusions in part from related forms. From San Diego County there are available numbers of Crotalus confluentus oreganus and Crotalus ruber, collecte(;i at various times throughout the year, all months being represented. From these we are able to plot ap- proximate growth curves and draw the following conclusions: 1. Young rattlesnakes are born in the early autumn, in most areas in September or early October. 2. Females go into their third hibernation (at an age of about 26 months), carrying developing eggs; fertilization occurs in the spring and the first brood is born about the time the mother snake has completed her third year of life. 106 229 M 85 242 127 392 841 (U) ~ I C C -P •H -H C OU%^ Lr^c-a^|<^^^^co|^^u^«5|^-C^lW^|c>r<)C-]'^U;^'-^C^JC-^r^C-0>|3^^^ 1 ■p o n! 10 o tn -H -P M ^< o o c >o o c ; rArOf>-)n-lcM|rHrv^^|^COC^CJNOsO|cMC>-lJ->|'^CMO|MCM O | O O •H -p o to o r-l Q 3 ?-< '•■) <^ 0) 0) O^OvO c 3 Osrnc 3 cr^u^cMrHmlor^H^^-"^^|^^-ovHoc5l^^c^JOO|o^-o I ro ^1 01 !t) (1) tfl a > ha 1' -a: frj Tt, 6 6r:<6AA64y-^c6^^^x><^rr■^rn^co^^cT^^^\o';:^^ c- 11) a) « ft ^ o o o c D rnU>< D mxi>c ncM o o unHctncm-o^o o -il-k^o t>j(>-0 C3|0 !>-o o u tn j^ > fcOPD . • • • • ^ rHrHrH ^^^r-{r^r-lr--: Cd-MS OO !>- ■^-^OOD jNOD cr- OCOO rH ?\Oli^ rH rH rH O 't^ ^ rH CU rH rH rH rH rH rH ^ GJ ^o;^ •H U V :0 til h •3cen o aM>- O l^^a^ £) rOrH rH rH CM CM CM rHrH rH rH rHO rH r^ rH r-{ f-A r-l O UNrH H rH CM ^CO C7- r-< ,-< •-< T3 C~-^ ^O U^ CM r-i r-l t-l f-< r-< r-l rH CM D- ■H > 1' 4-> i^?a O ONC-- ,-{ rH -O (MOD U^rH O vO--0 UN O C^CJ^ fV-iUM>. X> ■^ rr rH H O rH rH rH rH rH oco o 't CM O rH ^ UVXJ^JD 1^^ rH CM rH CM ro-* rH hj O hO trt ?H •>-l H iH 01 03 n cent alss a t3 03 +J rH 0 3 * crt rH O 03 g p M 00 \r\0 C^CM O UNm[> CM^O rH rocvl O CM rH O rH O CM CM O O CM O O o o o H rH rH rH rH 0) H ''3 C'^ s 1 tH •p ,o o m c o r-t> -C^O c -C--0 -^ ■CN.O «* ooc- -^ l>-I> -i>.o 'd ' O'l- C > O-O !> -rv-iO C^ O u P o > eii o H a) E o o O CM r- ( o ■* u- >^^ r^ ■ICM^ ^ lr\-o^ ) yDO u > cmun-^ 1- CM CM C ) rH CM O CD O (U ■H P M O rH ^ ^:? 3 rjj a. 03 Vh tS "J S o * a O rH ■- ( rH rnc il r-iOrX 3 C^sO U >fV>o C- -X) i^C rH r > rocj^a H rH D (-030^ D^ r^r H CM '^'-l CT^ r-l 3 ^H ID S c5 r^ 1 03 OCTnC ^a^o^c r- f>OC r ocjnc ^ o^oc ^C7^0^C -nOCJnC T>CTNaNC 1 1 1 rocjNC 1 1 1 r>loCTNCT- 1 1 1 1 r-l O 03 IS) a •H rH -.1 o 1 1 oo c P>-00 C 1 1 ) o o < ^ o rH c DO O C 0 O -O-l 300 C 3 OOC D ONO r DO OC H CM 1-^^ DOO C t LCNO C D OOC D O O C 0 r-IOtr D O O O nx)X>oo o 1 0> 03 rt I O ir\ o > •H 03 3 r^ This appears to be the typical life cycle, but there may be deviations. There seem to be well authenticated in- stances of snakes mating in the summer; and in the late spring some young have been found which were so small in size as to lead to the suspicion that they could not have been born in the autumn, but probably appeared subsequent to the hibernating season. Of course these statements have reference to snakes under natural conditions; the effect of captivity on growth and development is so material that it is doubtful whether observa- tions on captive specimens are of much value in determining growth and maturity curves. Fertility Statistics - Platteville Series; Returning to the Platteville series of conf luentus . we find the smallest female with developing eggs to be 588 mm. long, and that egg bearing continues up to the largest size reached by the females (863 mm.) The smallest true adults may therefore be considered to be about 580 mm. in length; how- ever it will be noted from the sixth column in Table 2, shewing gravid females in per cent of total females, that comparatively few of these small females are fertile, the figure averaging less than 20 per cent up to 64.O mm. At this point there is a sharp increase in fertility to somewhat in excess of 50 per cent. Thus, we may conclude that the length-range of 580 to 64.0 mm. contains some large adolescents as well as some small adults; this is probably the range in which the large year-old snakes overlap the small two-year-olds. Above 64.0 mm. the fer- tility rate Increases rapidly until it approaches 100 per cent in the largest sizes, as illustrated in Fig. 1. It is clear that with gravid females well over 50 per cent in all these larger sizes, it is proven (as would be expected) that young are normally borne annually. The distribution of gravid females amongst the size groups is Indicated in the third coliunn of Table 2, and by the solid line in Fig. 2, in which the results have been general- ized by smoothing out the fluctuations produced by the indivi- dual points. It will be observed that there is first a minor peak at 65O mm. followed by a principal peak at 750 mm. It is presumed that the first peak is the result of an advance to the adult stage of a group of adolescents, slightly more than two years old, while the second peak comprises the females which have reached full adult growth; the latter probably includes individuals from a number of successive annual classes.* The distribution of gravid females differs from that of total fe- males primarily in the sharpness of the two peaks; the spread of total females indicating a more even distribution. This would tend to confirm the presumed overlapping of adolescents one year old with the lagging adults of the preceding year. Columns 7 to 11 present the egg statistics in terms of total eggs, and minimum, maximum, and average eggs per gra- vid female; also the average eggs per female, whether or not gravid. The same data are set forth completely in Table 3. It will be noted that the eggs comprising a clutch vary in ♦There are no further peaks because the individual dispersion in size amongst the adults more than counterbalances the grad- ually increasing average size of successive age-classes. (16) 550 600 650 700 750 SIZE IN MILLIMETERS 800 850 900 lU 9 8 7 h- 6 Z UJ O 5 a. a. 4 Fig jre 'i 1 c )IST FF RIBl MAI JTIC FS )N C AN )F C .RAN /ID / '"N V C.c .conf lueni US — -Plal tevil le S( Ties / A \ \ GR EG WID FEK lALE s I V \ ••BM M^ y \ \ / r^ \ /! v \ \ / ' -\ \ > // ^v-> \ \ ,/ 1 t / \: s On ^ * o 0) § -J o o + CM rH CM i-t 1-1 r-i r-l CM •H CM (>- .-1 H r-t 1-1 r-t rH H CMrH CM r-1 •=t- CM ro CM rH rH ro rH on r^ rvi CM on rOH rH rO CM H rH CM CM nOrn rr^^ ■* rH rH rH rH ro CM CM rH rH rH rH rH CM H o rH O rH rH rH CM rH rH CM r-i rOrH CM •o rH C3N H rH rHrH rH H rH CM rOrH rH CM H CO rH •-i f^CM H CM rH rH CM -H c^ r^ CMrH 'd- ^o rH rH r-t ro \r\ r-i rH CM ■* (M CM 1 1 1 o o o ro ONO VrNU>\0 1 1 1 oo o rH(M f^ CTsCTNON 1 1 1 o oo ^ J3-0 1 1 1 O OO c^oo o ■O-O'O cj\oa> 1 1 1 o o o O rH CM c^r--o- OCnOn 1 1 1 OO o 1 1 1 o oo o i>--o oc3>a> 1 1 1 o o o CJNC7^CT> 1 1 1 O o O CM ro-* CTnOs 1 1 O O lf^^O coro rH o H •urai j.avnon JO sehxC-^zTS - number from L, to 21, and that there is a definite tendency to increase in nxamber per clutch as the size of the mother in- creases, from about 6 in the young mothers, to 15 or 16 in those of 800 to 820 mm. Beyond this a slight reduction is evident, although this cannot be considered definitely proven because of the small number of large fertile females available. The coefficient of correlation betv/een the size of the mother and the size of the brood is found to be H-0.711±0.039. The total egg production curve (dash line. Fig. 2) is slightly different from that showing the distribution of gravid females because of the effect of larger broods in the larger snakes (column 12, Table 2). Taking this group of confluentus as a whole we find that there were 1767 eggs in M9 females, with an average of 11.85±0.18 per batch. In the same size-classes there were 93 females without eggs, so that the eggs per individual of the total female population averaged 7.3. However, as has been pointed out, probably some of the 580-6^0 mm. snakes are only over-large adolescents. If v/e eliminate this group and the eggs which they produce, we find that the average eggs per fe- male is 10.3, which should be taken as indicating the produc- tivity of the fully adult female population, after allowing for infertile individuals. Birth Rate - Platteville Series: As 1767 eggs are produced by a population of 8;^1 snakes of both sexes and all sizes, including juveniles (which latter at the next bearing season will be adolescents), a birth rate of 2.10 is indicated. Hovfever, this assumes that the population in the late siumner iftill be of the same compara- tive composition as exists at the hibernating season; that there will be no selective losses. My experience would lead me to expect a high loss amongst the juveniles,* as compared with other elements of the population; this would tend to in- crease the apparent birth rate or ratio of young to the then existing population. Of course the number of eggs found in the females does not give a true picture of the live births to be expected. Examination of the eggs discloses a number v/hich are much smaller than their fellows of the same batch; some of these no doubt are defective. Broods born in captivity often include infertile eggs, distorted embryos, and deformed young snakes born dead or which succumb soon after birth. Vftiile these losses are probably of greater magnitude amongst captive snakes than are to be expected in nature, still there must be a con- siderable reduction from the egg count to live young. *Juveniles are more careless of concealment than adults and are preyed upon by various animals which could not successfully cope with an adult. See Klauber: Bull. Zool.Soc .San Diego, No. 8, p. 59, 1931. (19) We might hope to have some Indication of the popula- tion distribution which exists in early autumn, when the young are born, in the proportion of juveniles that are present (or at least captured) at the season of hibernation. We find the ratio of juveniles to adolescents and adults to be 229 to 612, or 0.374. Out of a total population of 612 existing in Septem- ber the gravid females probably niimbered II4. This figure is reached by eliminating the females which, as of the succeeding hibernating season, are carrying eggs for the first time. This gives a birth rate (based on juveniles still alive and collect- ed at hibernation) of only 2 per brood, which is decidedly too low. We are therefore forced to conclude that there is a heavy mortality of Juveniles between birth and hibernation, or that there is some selective effect tending to reduce the proportion of juveniles caught when the dens are raided in the autumn and spring. Probably both factors are effective. It is not Impos- sible that juveniles may favor individual retreats to a greater extent than adults, owing to inability to find the dens or the difficulty of traveling to them. If we allow an average loss of three per brood for Infertile eggs and defective young, we have 9 live births per brood and a birth rate of 1.65 young per unit of total popula- tion at the time of birth. Thus, for a short time in the late summer, the rattlesnake population is more than doubled, but the high mortality amongst the juveniles must soon cut this ratio extensively, so that by the time the hibernation season is at hand the juveniles already number less than the adult population, although probably not quite to the extent indicat- ed by the juvenile proportion of 26.7 per cent found in the Platteville series. This figure of 26.7 per cent is closely approximated in each of the four separate major catches which constitute the Platteville series, so that it may be taken as a true index of the number of juveniles that stray outside of the prairie dog holes. Mr. Perkins informs me that about the same relative proportion of adults as juveniles escaped during his raids. Selective losses are less important amongst the young in the spring season than in the autumn. If we assume that all of the snakes betv/een 4OO and 579 mm. were born the previous year, and that half of those between 580 and 639 mm. (this being the area of overlapping between the one- and two- year olds) represent last year's broods, we find that these adolescents represent 26.5 per cent of the population exclu- sive of juveniles; in other \\ords, the adolescents have the same ratio to the older classes that the juveniles have to those of an age greater than themselves. This would indicate that the young hold their own in the spring. Statistics of Jackley Series; The Jackley series of C.c.confluentus from the vi- cinity of Pierre, South Dakota, also furnishes some data on the size of egg clutches; but not on birth rate or proportion of gravid females, for this series does not represent an aver- age sample of the snake population, a full quota of juveniles . not having been collected. There are 673 snakes in the series; of these there are 107 gravid females, containing 1151 eggs. The eggs per clutch average 10.76±0.21 compared with 11.85^:0.18 in the (20) Platteville series. The clutches vary in size from U to 20. The smallest female containing developing eggs was 683 mm. com- pared with 588 mm. in the Colorado specimens. The maximum was 1020 mm., but this was a rather unusual specimen, the next largest being 950 mm. The largest Platteville female was 863 mm. Thus, it will be seen that the South Dakota adult females run about 100 mm. longer than those from Colorado j however, the pro- ductivity per female does not show an increase in South Dakota over Colorado, notwithstanding the correlation between size of mother and number of eggs within each series. This correlation in the South Dakota series is + 0.704i0.0A6 compared with the figure of +0.711± 0.039 found in the Platteville group. It should be understood that these statistics on both groups are made from counts of developing eggs. These are rela- tively large and yellow in color. Interspersed with them are frequently found white eggs of much smaller size which are not coiinted. Also, a few females are found (some clearly immature) which have only small white eggs of this type, presumably for development at some future time. These small eggs number up to 25 in a single female. They cannot be counted with great accu- racy, and as neither their time of development nor the propor- tion that will come to maturity is known, they have been omitted entirely from consideration in the statistics herein discussed. At the time of hibernation the developing eggs vary considerably in size, some typical clutches in the South Dakota series average in mm. as follows: 32 x 17, 28 x 15, 3A x 16^, 3-4i X 15i, 30 X 17^, 27 x U, 37 x 18, 30 x 14, 31 x 13, 21 x 10, 15 X 7|, 36^ X I6i, 3li x 13. There were altogether 121 eggs in these measured clutches. These figures are to be considered only approximate as the eggs are extremely soft and flexible, hence the dimensions depend somewhat on the method of support. There is considerable variation within each clutch; for example, a clutch averaging 32 x 17 imn. varied from 36 x 19 mm. to 27 X 14. mm. The undeveloped eggs, which are white in color, vary in all sizes from mere pinheads up to about 5 mm. x 3 mm., and their segregation from the developing eggs is rarely difficult. Egg Growth in Hibernation; It should be noted that the fall catches were made early in October, and that the snakes issued from hibernation in mid-April. However, in each case actual preservation took place about a month later, in November or May. Measurenents of a considerable number of the eggs in the Platteville series, in- dicate that there is little growth during the hibernating season, as is indicated by the follov.ing figures: Length mm. W idth mm. Max. Min. Aver. Max. Min. Aver 29 18 23.2 17 8 12.1 28| 18^ 24.5 151 8 12.8 Autiimn -------- Spring -------- The growth In each dimension during the hibernating season is thus found to average about 6 per cent. In making these measurements only normal, developing eggs were considered; the small, white non-developing eggs v.ere omitted. (21) o o o u a u o CO <0 si o +> (S m 60 (0 • C^CM *^ »0^^0 »co Ol rH rH rHvOO »C^Tl- C^"\rH X> rH»t \r\sO !>-■* o 4 r-l CM rH r^ r-l r-l r-l rA rH rH r-l rH •*CM0O r-i rnvi> -o »t O^t^OO CMH rH rH ^ O f O rH^ CM rH rH r-l O rH r-l CM UN H O iH H ro rn rH r-l OJ CM rH r-l W H r^ rH CM CM O r-l r-l r-l CM rn CM O rH rH CM CM H CO CM f-1 rH rnrH CM vO O r-l r-l C^ r^ n^r^ lr\ r-l r-l t^ ^ r-l rH r-l vO l>~rr\^ •* rH UN M i-l r-l U^ rHOO ro CM CM iH i-H CM CM «t rH mUNrH ON r-l CM rn •d »H rH r-l CO o ro CM roo rH •* rH ro^ r-l CM c^ o M rH rH CM rn CM rH rH rH <^0 H ■d- rHCM r-l Lr\ r-l t>- c r-l CM rH ro ^ •H »H rH CM CM O a^rn CM CM 1^ r-l rH r-i UN. (U r-l rH CM rn ^ o •O^O vO X) rH 't CM UN !2! r^ rH r-l rO "^ (y> \r\ mvr\ C^xr\rn U\ CMU^ rH CJ r-l C7N rH H ro UN CO rr^f-i CM CM ■* X) r^ rH CM r-l r^ CJ o r^ r-l CM ■* c~~ t^rH roCM ■"iJ-MD nr f^H ro CM CM CM r^ l>- r-l m vO r-l CM CM lr\rn rornco X> rA CM CM r-l r^ r^ r^ rA Oi nn rn U\ rn Hrn rH CM CM 'I- CM X)CM r^\r\r-l rH CM rn m ^o rn "t r^ r-l CM CM r^ f»1H CJNCM r-l Cn r-l ^ CM ro rn ■>t r^ rH 00 rH rn r-l rH rn CM CM r-l >-i r^ r-l rH r-l m ra o ID (D (0 +J 3 rH 31 C C 3 3 3 C 3 g> 01 m rt H 3 3 •H 3 n 3 +J C 0) rH ■H llj o n (D 0) rt ,c! J- ,rH O +J ID O "lO ^^ D< m -p T) 3 U t-i nS U 0) « (0 rH in (fl « in Vl 3 C! , ■P tt) ^^ (U -H OJ •H T) a) rt C rt ■p P< ■H H o a o o( 0) +J o *» rH , ■p ^ rt O U H 0} rH 0) -H o 3) M 01 H 3 ^ o ^ 3 O •p J4 o a c a rH O D 0) M ti >>u a ■r< rH -H rJ EH d 01 o 13 *J 3 3 O 0) o n +J ■H B a) OlrH U ID . rH a) -rt o • • • • • . +> O s o a OOO odo O d oo p< ^^ a Eh O O o o a g oo CJ O O OOO OOO oo MM I have previously mentioned the fact that the small- est fertile female In the Plattevllle series, has a length of 588 mm. and that the range, 580 - 64O mm. is presumed to con- tain half adolescents and half young adults, the faster growing members of the former class having caught up with the slower members of the latter. Attention should be called to the fact that these figures, 588 and 6^0 mm., lave reference to young adult females at the time of hibernatirn; when their first broods are born in the early autumn, these Individuals will have Increased somewhat in length so that their differences from the average size of the adult females will be somewhat reduced. From certain rattle studies I would estimate this Increase at not less than 120 mm. per snake, so that we may presume the smallest females to have a length of about 700 mm. at the time of the birth of their first broods. This refers only to the Plattevllle series of C.c.confluentus as found in east-central Colorado. In the Dakotas and Montana they will be about 100 mm. longer; and near Winslow, Arizona, where they have differentiated Into the stunted subspecies, C.c. nuntius. they will be about 250 mm. shorter than the Colorado specimens. Sizes of Rattlesnake Broods - Other Species; For comparative purposes such data as are available on sizes of broods in C.c.confluentus in addition to the Platte- vllle series, and in other species are given in Table U- There are represented 515 egg batches or broods with an average of 9.9 young per brood. Obviously, these data are too fragmentary to warrant extensive conclusions, since comparatively few broods are available in most species. The larger species (atrox, ruber, etc.) are probably not fairly represented, as the largest fe- males do not ordinarily appear in collections; thus we may expect that their broods will average somewhat larger than the tabular figures would indicate. The few broods of duris- sus and basiliscus suggest that these large rattlers are pro- bably prolific. Smaller species of rattlers, as for Instance, lepl- dus and triseriatus. have fewer young than the larger species. Stunted subspecies, such as C.c. nuntius. show the same tend- ency. Three batches from the stunted oreganus on Coronados Islands, averaged only 3 eggs. The embryos of these small species are both longer and heavier in proportion to the dimen- sions of the mother than are the embryos of the larger species; thus, there are physical limitations preventing more numerous broods in the smaller species. This fact is quite evident upon a visual inspection of full term embryos and parents, and is corroborated by measurements. Conclusions; 1. The smallest fertile females in C.c. con- fluentus from Colorado are about 585 mm. long. At 64.0 mm. "there is a sharp increase in fertility, which tends toward 100 per cent above 700 mm. These lengths are determined about 6 months prior to the birth of the initial broods. 2. The jgg production curve of C.c. confluentus reaches a maximum at 750 mm., ov;ing to the high pro- portion of adult females of approximately this length. (23) 3 . Broods of C . c . conf luentus vary from about 4 to 21. There is a definite increase in number of young with the size of the mother, the coefficient of correlation being about + 0.7. - e eggs average about 11 per batch. 4.. The birth rate of a rattlesnake population is about 1.65, but there is a considerable loss of juveniles between birth and hibernation, so that when the latter season arrives the juveniles number only one-quarter of the population. 5. The eggs seem to increase about 6 per cent in linear dimensions during the hibernating season. 6. Fragmentary data from species other than C « c . conf luentus indicate that fully adult rattle- snakes normally give birth to from 3 to 20 young, the average being 10 or 11. Smaller species, such as triseriatus and lepidus tend to have fewer young, usu- ally numbering 4- to 6. (24) cAc 'cme Bookbinding Co., Inc. tOO Camorid^e Si. Chartestown, MA 02129 3 2044 093 361 160