University of Kansas Publications MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publica- tions, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extra- mural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publica- tions by addressing the Exchange Librarian, University of Kan- sas Library, Lawrence, Kansas 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, University of Kansas, Lawrence, Kansas 66045. Eaiior: Richard F. Johnston PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS S~0- ifo \ rc-T^ • 2OOL OCCASIONAL PAPERS of the '976 MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 56, PAGES 1-8 MAY 26, 1976 EVOLUTION IN THE HOUSE SPARROW, V. COVARIATION OF SKULL AND HINDLIMB SIZES By Richard F. Johnston1 Big animals, as the sage said, have big parts, and small ones small parts. Lying back of this is an over-all morphologic covaria- tion that is familiar to everyone. It is nevertheless not so clear how far such covariation can be detected. One cannot readily predict how sizes of different bodily parts may covary within a species, that is, whether proportional relationships can cause significant covariation. Yet, certain bodily subregions might be expected to vaiy together because they are somehow functionally related. For instance, since sizes of the skull, beak, and hindlimbs of seed-eating birds are associated in how the birds get to, catch, and husk seeds for food, we might expect skull and hindlirnb sizes to have tighter covariation than either has with other bodily subregions. In fact, this generally appears to be true, and, using such food-getting morphology, we should be able to define a part of the food niche of a species, a relatively subtle thing. The task is partly to char- acterize for one or more populations the sizes, variances, and intercharacter correlations of the critical variables. Such characterizations for even a few such variables have al- ready provided instructive insights for students of niche ecology, interspecific and intraspecific competition, and niche partitioning (Van Valen, 1965; Selander, 1966; Fretwell, 1972; Rothstein, 1973; Cody, 1974; Hamilton, 1974). Thus, we know that restricted vari- ances in beak sizes in birds are found in populations subject to heavy pressure from interspecific food competitors, and large variances in those populations not subject to, or released from. 1 Museum of Natural History and Department of Systematica and Ecology, The University of Kansas, Lawrence, 66045. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY interspecific competition. Such apparent morphologic sensitivity is striking, and suggests that even more subtle ambiences of char- acter covariation may be detectable. The present study looks at degree of covariation between bill, skull, and hindlimb skeletal sizes in adult and first-year house sparrows Passer domesticus from one population in Manhattan, Kansas. Univariate assessments of correlations between pairs of size variables are compared to canonical correlations of the same sets of variables. The hypothesis is that adults will show tighter covariation than the subadults. Materials and Methods The specimens used are 20 adult male, and 38 subadult male, house sparrow skeletons, taken in October, 1971, from one flock at the dairy husbandly barns on the campus of Kansas State Univer- sity, Manhattan, Riley Co., Kansas. The adult birds were at least 15 months old. The subadult birds were hatched earlier in 1971; all had completed their postjuvenile molts but still had some re- gions of single-layered bone in their crania, and were roughly 4 to 6 months old. The extent of the temporal heterogeneity in the composition of these samples can be gauged as follows. Some 38 birds, all the subadults, were in their first year of life and had not lived through their first wintertime. Then, if we assume demographic charac- teristics of house sparrows as given by Summers-Smith (1963), half (10) the adults were in their second year of life and had sur- vived one wintertime period; half of the remainder, five, were in their third year of life and had seen two winters; and so forth, 2 or 3 were in their fourth year of life, 1 or 2 in their fifth, and perhaps 1 was as old as six years. Assumptions (cf. Lack, 1968; Fretwell, 1972) concerning winter as a time of major selective mortality loss in sparrow populations are of some consequence to the arguments developed here, and the important assumption is that the sample unexposed to wintertime stresses is different in its morphometries from the sample exposed to at least one wintertime period. The variables employed are skull, premaxilla, dentary, and mandible lengths, skull width, and femur, tibiotarsus, and tarso- metatarsus lengths. Measurements were taken with dial calipers reading to 0.1 mm, and a summary of statistics concerning the sample measurements is presented in Table 1. The analysis relies in part on assessing univariate correlation between pairs of variables. Correlation can be increased by high covariance, by low variance, or by both these phenomena acting together, a point worth noting when looking at correlations for the subadult specimen set, in which there is high character-pair co- 0.34 14.6-15.9 15.2 0.36 14.4-16.0 0.51 28.9-31.0 29.6 0.57 28.4-31.0 0.25 5.7- 6.8 6.1 0.25 5.4- 6.6 0.55 18.9-21.7 20.1 0.48 19.0-21.1 0.52 16.1-18.7 17.7 0.45 16.5-18.5 1.01 25.2-29.1 27.8 0.74 26.3-29.3 EVOLUTION IN THE HOUSE SPARROW 3 Table 1. — Standard Statistics for the Variables of the Head and Hindlimb In Two Samples of House Sparrows from Manhattan, Kansas. Adults (N=20) Subadults (N=38) X s Range X s Range PremaxiJla 6sT 0.25 6.4- 7.5 6.8 0.23 6.1- 7.2 Skull width 15.3 Skull length 29.8 Dentary 6.2 Mandible 20.3 Femur 17.6 Tihiotarsus 27.7 Tarsometatarsals 18.8 0.81 17.0-19.9 19.0 0.53 17.8-20.2 variance probably due to the specimens coming from a single age cohort. Statistical computations were performed at The University of Kansas Computation Center on a Honeywell 635 digital com- puter. Canonical correlation analyses were performed via a time- share program (CANCORR) written by Peter Neely. Results Intercharacter correlations of the skull variables. — Table 2 is a summary of the relationships between the five skull variables. Pre- maxilla, dentary, and mandible lengths are measures of bill length, one of the most important variables affecting foraging by house sparrows. The dimensions of the skull, both length and width, can be taken as assessing the surface area available for attachment of muscles that open and close the bill and which are therefore re- sponsible in catching and husking seeds. Table 2. — Correlations1 Among Skull Variables in Two Age Classes of Male House Sparrows from One Population in Eastern Kansas. Adults; N=20 Pre- maxilla Skull Skull Dentary Mandible Length Width Length Length Length ^ Premaxilla Length .438 .755 .702 .741 fc Skull Width _ .312 .430 .352 .482 | Skull Length .625 .615 .786 .864 1 Dentary Length .804 .296 .619 .728 | Mandible Length .661 .288 .619 .707 1 Product-moment coefficients exceeding 0.444 for adults, and 0.323 for sub- adults, show a significant correlation at P^0.05, and are set in italics. For both age classes there are seven significant correlation co- efficients. It is moot whether adults or subadults show a more meaningful set of intercharacter relationships, and the case for ap- 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY parent functional covariation here could be made as well with subadults as with adults. It is clear that there are differences be- tween the two sets of specimens, but whether this reflects the fact that the adult birds have lived through at least 15 months during which selection for efficient feeding must have occurred is not possible to detect. Inter character correlations of skull and leg variables. — Looking at the relationships between skull characters and the hind limb bones (Tables 3 and 4), the picture becomes somewhat obscure — adults show only three of 15 possible correlations between skull and leg to be reliable, but the subadults have six significant co- efficients. The data thus show a higher degree of covariation between head and leg in the subadult birds if covariation is assessed one character at a time. Canonical correlations between head and leg variables of both age classes are shown in Table 5. The subadult sample has a low first canonical correlation coefficient (CCC I), the chi-square value computed for which fails to achieve statistical significance (.5