There are diverse relations among ecological, morphological, and life-history traits in North American woodpeckers (family Picidae).Within the family as a whole, clutch size does not correlate with body size. However, clutch size increases with body size within the genus Melanerpes and decreases with size in Picoides. In the family as a whole, species that specialize on wood-boring larvae have small clutches. Such species use pecking as a major foraging technique, and pecking is associated with a wide suite of morphological specializations, including relatively wide first ribs, long pygostyle disks, short tibia, wide crania, wide maxillae, long mandibular symphyses, less cranial kinesis, and greater culmen sexual dimorphism. Hence, these morphological characters also correlate with clutch size, in two cases (length of the mandibular symphysis and cranial kinesis) even after controlling for both body size and generic effects. The observed correlations between clutch size and morphology are probably the result of dietary influences. These correlations, however, may at least in part be due to morphological constraints. Morphological design may thus constrain life-history evolution at the intrafamilial and intrageneric levels as well as at higher taxonomic levels. Received 8 September 1986, accepted 20 May 1987.

Hastings Reservation and Museum of Vertebrate Zoology, University of California, Star Route Box 80, Carmel Valley, California 93924 USA NUMEROUS studies have pointed out strong classwide correlations between life-history traits (LHTs) and body size (e.g. mammals: Blueweiss et al. 1978, Western 1979, Tuomi 1980, Millar and Zammuto 1983, Stearns 1983; birds: Ama- don 1943, Lack 1968, Rahn et al. 1975; reptiles: Stearns 1984). These studies suggest that phy- logenetic constraints, as reflected by differences in body size, may be important in determining life-history variation. Further, covariation in LHTs may be an indirect consequence of selec- tion on size alone rather than a direct adaptive shift to environmental conditions or ecological factors (Stearns 1983, 1984). These results also raise the possibility that analogous types of morphological constraints may be important in shaping LHTs at low taxonomic levels as well, such as on an intrafamilial or even intraspecific basis. The existence of such low-level interac- tions would indicate that correlations deter- mined from broad-scale, classwide analyses may be incomplete or misleading, and thus would have important implications to the interpreta- tion of interspecific variation in life history. Nonetheless, there have been few studies of the ecological correlates of clutch size in closely related taxa, and virtually none of morpholog- ical correlates, with the exception of correla- tions with body size. I analyzed interspecific patterns of clutch-size variation in North Amer- ican woodpeckers (family Picidae) with respect to dietary and morphological characters in this family. My goal was to identify the correlations among these three sets of traits and to discuss the possible causal relationships producing the observed patterns. Although comparative data on LHTs other than clutch size are not currently available for woodpeckers, the extensive back- ground of ecomorphological studies on this tax- on make it an excellent one for such an ex- ploratory analysis. MATERIALS AND METHODS North American woodpeckers, excluding the vir- tually extinct Ivory-billed Woodpecker (Campephilus principalis), include 21 species in 5 genera. For all species except Acorn Woodpeckers (Melanerpes for- micivorus), data on clutch size were taken from mu- seum collections. Only data from complete clutches, as indicated by the incubation status listed by the original collector, were used. The locality for each clutch was recorded, and the latitude was determined to the nearest degree. Because two or more females may contribute to nests in the Acorn Woodpecker (Koenig and Pitelka 1979, Mumme et aL 1983), I used clutch sizes only from 90 nests of single females de- termined as part of an ongoing field study of this species at Hastings Reservation in Monterey Co., Cal- ifornia (Koenig and Mumme 1987). TABLE 1. Mean clutch size of North American woodpeckers. Sample sizes are given in parentheses. Mean clutch size adjusted for Species Mean clutch size + SD latitude Body mass a Melanerpes lewis 5.88 ñ 1.19 (51) 5.70 97.7 (7) M. erythrocephalus 4.82 + 0.80 (71) 4.74 66.8 (3) M. forrnicivorus b 4.36 ñ 1.03 (90) 4.46 79.8 (12) M. uropygialis 3.75 ñ 0.93 (69) 4.13 62.3 (18) M. aurifrons 4.65 + 0.80 (55) 5.39 79.8 (4) M. carolinus 4.31 + 0.76 (61) 4.63 69.1 (5) Sphyrapicus varius 4.93 + 1.02 (46) 4.33 47.2 (2) S. nuchalis 4.76 + 1.09 (17) 4.57 50.8 (6) S. ruber 4.69 + 0.72 (42) 4.48 52.2 (8) S. thyroideus 5.67 + 0.85 (49) 5.56 52.3 (10) Picoides scalaris 3.81 + 0.76 (58) 4.30 31.6 (16) P. nuttallii 4.34 + 0.77 (88) 4.46 36.0 (10) P. pubescens 4.81 + 0.78 (205) 4.63 26.7 (20) P. villosus 3.93 + 0.72 (159) 3.75 63.4 (27) P. stricklandi 3.50 + 0.84 (6) 3.87 42.7 (5) P. borealis 3.68 + 0.69 (25) 4.08 47.3 (20) P. albolarvatus 4.35 + 0.73 (98) 4.31 55.8 (12) P. tridactylus 3.87 + 0.64 (15) 2.83 55.3 (3) P. arcticus 3.89 + 0.85 (27) 3.21 70.8 (4) Colaptes auratus 6.21 ñ 1.56 (467) 6.19 146.4 (20) Dryocopus pileatus 3.80 ñ 0.77 (92) 3.74 265.3 (4) Females taken September-March; mass in grams. Clutches of single females only (see text). Mean clutch size was determined for each species and then adjusted by an analysis of covariance con- trolling for the latitude at which each clutch was col- lected; only the adjusted values were used in subse- quent analyses. This procedure corrects for the latitudinal gradient in clutch size within the family (Koenig 1986) and results in estimates of mean clutch size independent of latitudinal differences in range. Morphological measurements generally were taken from previous authors, including Ridgway (1914, us- ing mean values for the most widespread subspecies), Butt (1930), Spring (1965), and Kirby (1980, in litt.). Measurements used were of females when possible, because females are directly responsible for laying the clutches whose size was of interest. Measurements from males and females were highly correlated, how- ever, for all characters for which I had data (corre- lations for femur length, pygostyle disk length, tibia length, culmen length, body mass, cranium width, and wing length all r 0.90, P < 0.001). Thus, anal- yses using measurements from males would yield similar or identical results. Body-mass data were taken from females collected during the nonbreeding season (September-March) and are from specimens in the Museum of Vertebrate Zoology. Only a single such specimen was found for the Red-cockaded Woodpecker (Picoides borealis), so I used body-mass data for this species from live birds caught between September and December in North Carolina (R. Repasky pets. comm.). Dietary information is from Beal (1911); egg mea- surements are from Bent (1939). Relative egg mass was estimated by the "egg-value" (length x breadth 2 / 1,000; Amadon 1943). This value was then divided by female body mass to yield relative egg volume and, in a separate calculation, multiplied by mean clutch size and divided by female body mass to derive the relative clutch volume for each species. Such ratio measures can be misleading. As the primary focus of the analyses performed here is on clutch size varia- tion, however, they are adequate to outline general patterns in these measures. Taxa used follow the A.O.U. (1983, 1985). In all, data on 1,791 clutches were used in the analyses. I use nonparametric Spearman rank tests for pairwise correlations and, in lieu of a comparable nonpara- metric multivariate technique, multiple regressions and ANCOVA for examining interrelationships of three or more variables. P values -< 0.05 (two-tailed for Spearman rank tests) are considered significant. Two potential shortcomings of these data should be noted. First, I have used data from a variety of sources. As a result, sample sizes vary and the data are undoubtedly of variable quality. I have preferred older sources to measurements of my own, however, because the data they contain invariably were gath- ered for reasons unrelated to the analyses performed here. Thus, the results are certainly unbiased. Second, although I have taken into account some of the geo- graphic variation in clutch size by controlling for latitude, I have ignored intraspecific variation in mor- phology. Although such patterns are of considerable T^BLE 2. Spearman rank correlations of clutch size with body size and culmen dimorphism in North American woodpeckers. a Clutch size adjusted for Relative egg Relative clutch Variable latitude volume volume Source All species (n = 21 species; 17 for femur length) Female body mass 0.17 - 0.83' * * - 0.60 * * This study Male femur length 0.16 -0.85*** -0.64** Kirby in litt. Female tarsus length 0.21 -0.77*** -0.51' Ridgway 1914 Culmen dimorphism b -0.51' 0.24 -0.47* Ridgway 1914 Genus Melanerpes (n = 6 species) Female body mass 0.70 -1.00'** 0.41 This study Male femur length 0.77 -0.99*** 0.43 This study Female tarsus length 0.83* -0.67 0.77 Ridgway 1914 Culmen dimorphism b -0.49 0.32 -0.83* Ridgway 1914 Genus Picoides (n = 9 species) Female body mass -0.70* -0.67* -0.93*** This study Male femur length -0.84** -0.65 -0.96*** This study Female tarsus length -0.57 -0.58 -0.87** Ridgway 1914 Culmen dimorphism b -0.13 0.13 -0.02 Ridgway 1914 * = P < 0.05, ** = P < 0.01, *** = P < 0.001. Expressed as the ratio of mean male to female culmen length. interest (e.g. James 1970), I have chosen to simplify the analyses by using mean morphological measure- ments from the most widespread subspecies when authors examined more than one race. RESULTS The body mass of extant North American woodpeckers spans an order of magnitude from the 26.7-g Downy Woodpecker (Picoides pubes- cens) to the 265.3-g Pileated Woodpecker (Dry- ocopus pileatus) (Table 1). Mean clutch sizes vary slightly less than twofold, from 3.50 in Strick- land's Woodpecker (P. stricklandi) to 6.21 in the Northern Flicker (Colaptes auratus); differences in clutch size among species are significant (Kruskal-Wallis 1-way ANOVA; X 2 = 765, df = 20, P < 0.001). Controlling for latitude, mean clutch size ranges from 2.83 in the Three-toed Woodpecker (P. tridactylus) to 6.19 in the North- ern Flicker. Clutch size, body size, and culmen dimorphism.-- In the family as a whole, clutch size is uncor- related with female body mass, male femur length, or female tarsus length, three indices of body size (Table 2). Both relative egg and clutch volume, however, are inversely and signifi- cantly correlated with body size. Thus, as in several other avian taxa (Rahn et al. 1975), larger species of woodpeckers lay relatively small eggs and have small relative clutch volumes (Table 2). A primary reason for the lack of any corre- lation between clutch size and body size in the family as a whole is that the two major genera Melanerpes and Picoides provide opposite clutch to body size relationships (Table 2). Within Me- lanerpes, clutch size increases with body size, significantly so using tarsus length, whereas within Picoides the reverse is true. Relative egg volume decreases with body size in both taxa, whereas relative clutch volume decreases with body size in Picoides but not in Melanerpes. Cor- relations with culmen dimorphism are not sig- nificant except for an inverse correlation with relative clutch volume in Melanerpes. Clutch size and morphology independent of body size.--I performed two sets of analyses to assess the relationship of a series of seven morpho- logical variables with clutch size while con- trolling for the effects of body size. First, I re- gressed clutch size on pairs of morphological characters. In each case the first character was the one whose influence on clutch size was di- rectly of interest, while the second was femur length, a good index of overall body size in woodpeckers (Kirby 1980). The significance of each variable as a predictor of clutch size was calculated while controlling for the other; hence, the effect of each morphological character in- dependent of body size was determined. This procedure avoids the difficulties that arise from the alternative of using ratios to standardize for TABLE 3. Clutch size and morphology in North American woodpeckers? Pygo- style- Mahdi- First-rib disk Tibia Cranium Maxilla bular Cranial width length length width width symphysis kinesis Multiple regression Variable c Standardized fi -1.05 -1.30 4.38 -1.39 -2.16 -1.11 0.92 F-value 26.2*** 5.7* 7.4* 6.0* 23.1'* 13.3'* 19.5'* Femur length Standardized fi 0.72 1.06 -4.39 1.17 1.91 0.71 0.43 F-value 12.4' * 3.8 7.5* 4.2 18.2' * 5.5* 4.5 Multiple R 2 0.65 0.42 0.48 0.43 0.74 0.63 0.74 Overall F-value 13.1'** 2.9 3.7 3.0 11.6'** 6.7* 9.9** Analysis of covariance (F-values) Variable 0.3 0.2 3.6 1.3 5.1 11.9 * 106.3 * * Femur length 0.0 0.0 4.4 0.4 1.7 4.9 4.3 Genus 5.5' 3.3 6.1 4.2 7.3' 12.6' 284.2'** Overall F-value 4.1' 2.3 4.8 3.1 5.9 10.7' 219.5'** n species 17 11 11 11 11 11 10 Source Kirby Burt Burt Burt Burt Burt Spring in litt. 1930 1930 1930 1930 1930 1965 Regressions are of clutch size adjusted for latitude on each pair of variables. All variables except first-rib width and cranial kinesis are for females. Femur length is used to control for body size (see text). For individual variables, df = 1, n - 3 where n is the number of species; for the overall F-value, df = 2, n - 3. For the analyses of covariance, the influence of the variable is assessed controlling for femur length and generic effects (see text). For the variable and femur length, df = 1, n - 1; for generic effects, df = 4, n 1; overall df = 6, n 1. b * = p < 0.05, ** = P < 0.01, *** = P < 0.001. ß "Variable" refers to the variable listed at the top of the table. body size (see Atchley 1978, Atchley and An- derson 1978, Blem 1984). Second, I performed analyses of covariance using the morphological variable of interest and femur length as covariates and genus as a factor. The purpose of these analyses was to examine the relationships between each variable and clutch size independent of both body size and generic effects; thus, the factor was controlled before consideration of the covariates. These tests thus correct both for biased taxonomic rep- resentation within the family (see Clutton-Brock and Harvey 1984) and for differences in the relationships of variables between genera. From the multiple regression, clutch size is significantly related to all seven characters tested, including first-rib width, pygostyle-disk length, tibia length, cranium width, maxilla width, length of the mandibular symphysis, and cranial kinesis (Table 3). Clutch size is also sig- nificantly related to femur length in four cases, despite the lack of significant correlations be- tween clutch size and body size alone (Table 2). The overall F-values are significant for four of the seven regressions, with the proportion of variance in interspecific mean clutch size ex- plained (multiple R 2) ranging from 42% to 74%. The analyses of covariance (Table 3) offer ad- ditional support of real relationships between two of these morphological variables and clutch size. Length of the mandibular symphysis and cranial kinesis are both significantly related to clutch size, while femur length is not signifi- cant in any of the seven tests. Generic effects are significant in four of the analyses. These results suggest that the significant effect of fe- mur length found in the regressions is due to intergeneric differences (see Table 2), and that at least some of the correlations between mor- phology and clutch size within the family are not based solely on either allometry or phylo- genetic differences between taxa. Correlation of clutch size and diet.--I correlated clutch size with five major components of the diet of woodpeckers as determined by Beal (1911). The results (Table 4) show a significant inverse correlation of clutch size with the per- centage of animal matter in the diet; the ap- parent reason for this is the significant inverse correlation between clutch size with the per- centage of the diet composed of beetles (Co- leoptera). No significant correlations exist be- tween clutch size and the percentage of the diet made up of fruit, mast, or ants (Formicidae), or between diet and relative egg or clutch volume. Within the genus Picoides there is also a signif- icant inverse correlation between clutch size and Coleoptera in the diet; in addition, species that eat more beetles have relatively smaller eggs. None of the correlations with diet are significant within the genus Melanerpes. In summary: (1) For all North American pi- cids and within the genus Picoides, species that eat wood-boring coleopteran beetle larvae have smaller clutches. (2) Larger species of Mela- nerpes lay larger clutches, while larger species of Picoides lay smaller clutches. Within the fam- ily as a whole, there is no relationship between clutch size and body size (Table 2); significant effects of femur length when analyzed in mul- tiple regressions disappear when genera are in- cluded in an ANCOVA and thus appear to be due to intergeneric effects (Table 3). (3) Species with large clutches tend to have less sexual di- morphism in culmen length, relatively narrow first ribs, short pygostyle disks, long tibia, nar- row crania, narrow maxillae, short mandibular symphyses, and greater cranial kinesis after controlling for body size. Only the last two vari- ables, however, are still related significantly to clutch size when further controlling for generic effects. DISCUSSION WHY THE CORRELATIONS BETWEEN CLUTCH SIZE AND MORPHOLOGY? Although body size correlates with clutch size in Melanerpes and Picoides, the opposite direc- tions within these two genera result in no over- all body-size effect within the family as a whole. As in mammals (Western 1979, Stearns 1983), larger Picoides lay smaller clutches. In contrast, as in most poikilotherms (e.g. Roff 1981), the reverse appears true in Melanerpes. Overall, larger species lay smaller eggs and have smaller total clutch volumes relative to their body mass, similar to the pattern found in other homeo- therms (Amadon 1943, Lack 1968, Rahn et al. 1975, Stearns 1983). Stearns (1983, 1984), on the basis of similar types of correlations between body size and life- history traits in mammals and reptiles, sug- TABLE 4. Spearman rank correlations of clutch size and diet in North American woodpeckers. Clutch size Relative Relative Variable adjusted egg clutch (n species) for latitude volume volume All species % animal (19) -0.47* 0.32 -0.06 % Coleoptera (15) -0.64** 0.29 -0.39 % Formicidae (16) 0.16 0.01 0.18 % fruit (16) 0.39 -0.38 0.07 % mast (16) 0.25 0.20 0.18 Genus Melanerpes % animal (6) 0.31 0.06 0.03 % Coleoptera (4) 0.40 0.80 0.40 Genus Picoides % animal (8) -0.67 -0.10 -0.38 % Coleoptera (6) -0.83* -0.94** -0.77 Dietary data is from Beal (1911). * = P < 0.05, ** = P < 0.01. gested that selection for size alone may be responsible for many widespread patterns in life-history evolution. [Specific conclusions from the latter study must be taken with caution (Dunham and Miles 1985, Hedges 1985, Vitt and Seigel 1985), but Stearns's general conclusion that phylogenetic constraints cannot be ignored when considering the evolution of life-history traits holds (Dunham and Miles1985).] Size may have been similarly important in influencing clutch-size evolution within these two wood- pecker genera as well. I do not know the reasons for the opposite relationships between clutch size and body size in these two genera. Western and Ssemakula (1982) concluded that a large proportion of vari- ance in life-history patterns both between mammals and birds and within orders of birds can be explained by differences in neural mass, metabolic rate, and body temperature. How any of these features vary within or between genera of woodpeckers is unknown. In any case, the opposing correlations between clutch size and body size found in the two largest genera ex- amined here suggests that design constraints such as size may act in quite different ways within lower taxa, and thus that trends found at or above the familial level may be misleading. Independent of body size, there is a diverse array of morphological characters correlating significantly with clutch size within the family. Although the interpretation of these correla- tions is not immediately apparent, a relatively straightforward explanation exists for many of the trends in Table 3, stemming from the cor- relation between clutch size and the proportion of beetles in the diet (Table 4). This fraction of the diet primarily represents wood-boring lar- vae, which are obtained by "pecking" (strikes or blows made in limbs and trunks of trees with the bill). This behavior has long been known to be associated with numerous morphological specializations. Hence, the tendency for clutch size to decrease with increasing specialization on wood-boring larvae and increasing reliance on pecking may explain the following corre- lations: First-rib width.--Shufeldt (1900) first pointed out that woodpeckers exhibit a pronounced broadening of the first thoracic rib, a phenom- enon whose functional significance was exam- ined comparatively by Kirby (1980). She con- cluded that species that excavate frequently while foraging have relatively wider ribs. The significance of this pattern is due to the greater musculature associated with the wider rib, sta- bilizing the base of the neck against bending, and thus allowing the generation of more force- ful blows during foraging (Kirby 1980). Wid- ening of the ribs thus enables woodpeckers to specialize on prey obtained by pecking. Such species have smaller clutch sizes, and thus the relative first-rib width correlates inversely with clutch size in the multiple regression analysis (Table 3). Pygostyle-disk length.--The pygostyle disk serves as a major point of attachment for the tail musculature (Burt 1930). The tail in wood- peckers is used as a prop (Thomson 1964); hence, the pygostyle disk is relatively longer in species that forage by pecking for larvae (Burt 1930). Thus, the pygostyle-disk length correlates in- versely with clutch size; that is, species that peck for food have relatively large pygostyle disks and small clutches. Tibia length.--Spring (1965) discussed the skeletal modifications for climbing and pecking in woodpeckers and concluded that shortening of the distal leg bones is a modification for the delivery of harder blows. He suggested that this shortening might increase the effective power of the muscles inserting on them, thereby in- creasing the hardness of blows made by the bird as it pulls its body toward the tree. Shorter distal bones also might allow the bird to hold itself relatively farther from the trunk, thereby aug- menting its capacity to generate blow momen- tum. Within the family as a whole, species with relatively shorter tibia are those that peck for food and have smaller clutch sizes. Cranial measurements.--Burt (1930) discussed a series of cranial adaptations of woodpeckers, including the width of the cranium, width of the maxillae, and length of the mandibular sym- physis. These morphological features all in- crease skull strength and thus the ability of the bird to inflict hard blows. In turn, pecking species have wider crania, wider maxillae, and longer mandibular symphyses. Within the fam- ily, all three of these characters correlate in- versely with clutch size, the last significantly so even when controlling for generic effects (Table 3). Cranial kinesis.--Cranial kinesis is the mech- anism that permits the upper bill to move in- dependently with respect to the brain case (Bock 1964, Spring 1965). Spring (1965) considered several functional hypotheses of differing de- grees of cranial kinesis in woodpeckers. He con- cluded that this phenomenon acts as a buffering mechanism, and that to deliver more forceful blows during pecking, kinesis would decline as the bones and ligaments tending to restrict cra- nial movement were used to steady the bill at the moment of impact. Following this reason- ing, the family-wide correlation between ki- nesis and clutch size follows from the tendency of species dependent on pecking to have re- duced cranial kinesis. Sexual dimorphism in culmen length.--Cabana et al. (1982) suggested, primarily on the basis of sexual-selection theory, that there should be a positive correlation between the degree of sexual dimorphism in body size and female re- productive effort. Their analyses showed a pos- itive correlation between sexual dimorphism and clutch mass in 114 species of nonpasserines due mostly to increasing clutch size in species with relatively large males. The correlation of both clutch size and relative clutch volume with sexual dimorphism in bill length in North American woodpeckers found here (Table 2) is in the opposite direction, and thus does not support Cabana et al.'s (1982) thesis. An alternative explanation for the pattern ob- served here stems from Selander's (1966) hy- pothesis of ecological release leading to diver- gence in foraging behavior and bill dimorphism under conditions of low competition. From this hypothesis the greater bill dimorphism in peck- ing species would be predicted if there are few competitors for wood-boring larvae. This lack of competition might then correlate with wood- boring larvae being a relatively stable resource, a situation correlating with high adult survi- vorship and small clutch size (Cody 1966, Stearns 1976). There are no data to determine unambigu- ously if there is relatively little competition for wood-boring larvae. Nonmigratory species of woodpeckers eat significantly more Coleoptera than migratory species, however (Mann-Whit- ney U-test, U = 9.5, n = 7, n2 = 8, P < 0.05). At the very least this indicates that wood-boring larvae are more stable on an annual basis than other resources used by members of this family. Besides a relative lack of competition leading to ecological release, an additional feature that might select for sexual dimorphism is if wood- boring larvae were relatively rare. A scarcity might increase sexual dimorphism in foraging behavior (Peters and Grubb 1983) and thus lead to morphological dimorphism. Additional life- history data are necessary to test this hypothesis critically. INTERPRETATION OF THE INTERCORRELATIONS A long history of ecomorphological studies in woodpeckers argues for the relationship be- tween morphology and diet being adaptive. How, though, can we interpret the correlation between clutch size and diet? One interpretation, discussed above for the relationship between diet and culmen dimor- phisre, is that wood-boring larvae may be more stable than other food resources or differ in availability in some other way so as to afford high adult survivorship and relatively low re- productive effort. Again, if wood-boring larvae were seasonally stable but relatively rare, this might further selection for reduced reproduc- tive effort. By this line of reasoning, the vari- ation in clutch size among species is due to the dietary differences among them, and the mor- phological correlates of clutch size are an in- direct result of the mutual dependence of mor- phology and clutch size on diet. An alternative interpretation is that the re- lationship between clutch size and diet may be, at least in part, a side effect of selection for the morphological specializations attendant to di- etary shifts. Because of the presumably func- tional relationship between diet and morphol- ogy, to the extent that the latter constrains clutch size, a correlation between clutch size and diet would emerge even if no functional relation- ship exists between them. Numerous workers have demonstrated that phylogenetic constraints, as reflected by differ- ences in body size, are important in determin- ing classwide variations in life-history variation (e.g. Western 1979; Stearns 1983, 1984). The cor- relations between clutch size and body size within the genera Picoides and Melanerpes found here suggest that size constraints may be im- portant within these lower taxa as well, albeit in opposite directions. In addition, the corre- lations between clutch size and morphology in- dependent of body size found within the family suggest the possibility that other types of mor- phological constraints may influence clutch size. Thus, clutch size within the family Picidae and within the genera Picoides and Melanerpes may at least in part be a secondary consequence of constraints imposed by size and by other mor- phological adaptations to particular foraging modes. I am unable to propose any likely scenario as to the exact nature of the constraints that might be involved. Certainly, no obvious relationship exists between any of the skeletal features pre- sented in Table 3 and organs involved directly with producing eggs. Nevertheless, virtually any morphological modification could constrain clutch size if associated muscle or skeletal de- velopment requires space or energy that might otherwise be devoted to egg production. Al- though such interactions are speculative, the same is true of the precise nature of the con- straint on fecundity imposed by large body size (e.g. Blueweiss et al. 1978; Western 1979; West- ern and Ssemakula 1982; Stearns 1983, 1984; Calder 1984). There is considerable interest in the inter- actions between ecology and morphology; in- deed, it is this interface with which the field of ecomorphology is concerned. However, corre- lations between life-history phenomena and either morphology, with the exception of body size, or feeding ecology have been virtually ig- nored. My analyses suggest that these major phenotypic features may covary in a coherent fashion both within families and within genera. Similar analyses on taxa for which more com- plete demographic data are available, such as primates and rodents, are likely to be particu- larly illuminating. Examination of intraspecific correlations between morphology and life-his- tory patterns would also be of considerable in- terest. ACKNOWLEDGMENTS I thank the museums and their curators who made the data used in these analyses available: the Amer- ican Museum of Natural History, California Academy of Sciences, Chicago Academy of Sciences, Delaware Museum of Natural History, Florida State Museum, Milwaukee County Public Museumß Moore Labora- tory of Zoology, Museum of Vertebrate Zoologyß Na- tional Museum of Natural History, San Bernardino County Museum, Santa Barbara Museum of Natural History, and Western Foundation of Vertebrate Zo- ology. V. C. Kirby and R. Repasky kindly allowed me to use unpublished data. J. Conley, N. Joste, R. Mumme, and P. Williams helped in various aspects of data collection, and the Department of Biology, Occidental College, graciously provided access to computer facilities. Various drafts of this manuscript were improved by the comments of L. Freed, H. Greeneß R. Mumme, K. Sullivan, P. Williams, R. Zink, and a host of anonymous reviewers. Financial support came from the Frank M. Chapman Memorial Fund and NSF grant BSR84-10809. LITERATURE CITED AMADON, D. 1943. Bird weights and egg weights. Auk 60: 221-234. AMERICAN ORNITHOLOGISTS' UNION. 1983. Check-list of North American birds, 6th ed. Washington, D.C., Amer. Ornithol. Union. 1985. Thirty-fifth supplement to the Amer- ican Ornithologists' Union check-list of North American birds. Auk 102: 680-686. ATCHLEY, W. R. 1978. Ratios, regression interceptsß and the scaling of data. Syst. Zool. 27: 78-83. --, & D. R. ANDERSON. 1978. Ratios and the sta- tistical analysis of biological data. Syst. Zool. 27: 71-78. BEAL, F. E. L. 1911. Food of the woodpeckers of the United States. U.S. Dept. Agr., Biol. Surv. Bull. 37. BENT, A.C. 1939. Life histories of North American woodpeckers. U.S. Natl. Mus. Bull. 174. BLEM, C. R. 1984. Ratios in avian physiology. Auk 101: 153-155. BLUEWEISS, L., H. FOX, V. KUDZMA, D. NAKASHIMA, R. PETERS, & S. SAMS. 1978. Relationships between body size and some life history parameters. Oeco- logia 37: 257-272. BocI, W.J. 1964. Kinetics of the avian skull. J. Mor- phol. 114: 1-42. BURT, W. H. 1930. Adaptive modifications in the woodpeckers. Univ. California Publ. Zool. 32: 455- 524. CABANA, G., g. EREWIN, g. H. PETERS, & L. RANDALL. 1982. The effect of sexual size dimorphism on variations in reproductive effort of birds and mammals. Amer. Natur. 120: 17-25. CALDER, W.A. 1984. Size, functionß and life history. Cambridge, Massachusetts, Harvard Univ. Press. CLUTTON-BROCK, T. H., & P. H. HARVEY. 1984. Com- parative approaches to investigating adaptation. Pp. 7-29 in Behavioural ecology: an evolutionary approachß 2nd ed. (J. R. Krebs and N. B. Davies, Eds.). Sunderland, Massachusettsß Sinauer. COD3/4, M. L. 1966. A general theory of clutch size. Evolution 20: 174-184. DUNHAM, A. E., & D. B. MILES. 1985. Patterns of covariation in life history traits of squamate rep- tiles: the effects of size and phylogeny reconsid- ered. Amer. Natur. 126: 231-257. HEDGES, S.B. 1985. The influence of size and phy- logeny on life history variation in reptiles: a re- sponse to Stearns. Amer. Natur. 126: 258-260. JAMES, F.C. 1970. Geographic size variation in birds and its relationship to climate. Ecology 51: 365- 390. KIRBY, V. C. 1980. An adaptive modification in the ribs of woodpeckers and piculets (Picidae). Auk 97: 521-532. KOENIG, W.D. 1986. Geographical ecology of clutch size variation in North American woodpeckers. Condor 88: 499-504. , & R. L. MUMME. 1987. Population ecology of the cooperatively breeding Acorn Woodpeck- er. Princeton, New Jersey, Princeton Univ. Press. ß & F. A. PITELKA. 1979. Relatedness and in- breeding avoidance: counterploys in the com- munally nesting Acorn Woodpecker. Science 206: 1103-1105. LACKß D. 1968. Ecological adaptations for breeding in birds. London, Methuen. MILLAR, J. S., & R. M. ZAMMUTO. 1983. Life histories of mammals: an analysis of life tables. Ecology 64: 631-635. MUMME, R. L., W. D. KOENIG, & F. A. PITELKA. 1983. Reproductive competition in the communal Acorn Woodpecker: sisters destroy each other's eggs. Nature 306: 583-584. PETERS, W. D., & T. C. GRUBB, JR. 1983. An experi- mental analysis of sex-specific foraging in the Downy Woodpecker Picoides pubescens. Ecology 64: 1437-1443. RAHN, H., C. V. PAGANELLI, & A. AR. 1975. Relation of avian egg weight to body weight. Auk 92: 750- 767. RIDGWAY, R. 1914. The birds of North and Middle America. VI. U.S. Natl. Mus. Bull. 50. ROFF, D. 1981. On being the right size. Amer. Natur. 118: 405-422. SELANDER, R.K. 1966. Sexual dimorphism and dif- ferential niche utilization in birds. Condor 68: 113-151. $1-IUFELDT, R.W. 1900. On the osteology of wood- peckers. Proc. Amer. Phil. Soc. 39: 578-622. SPRNC, L. W. 1965. Climbing and pecking adapta- tions in some North American woodpeckers. Condor 67: 457-488. STEARNS, $. C. 1976. Life history tactics: a review of the ideas. Q. Rev. Biol. 51: 3-47. ß 1983. The influence of size and phylogeny on patterns of covariation among life-history traits in mammals. Oikos 41: 173-187. ß 1984. The effects of size and phylogeny on patterns of covariation in the life history traits of lizards and snakes. Amer. Natur. 123: 56-72. TI-IolvisoN, A. J. 1964. A new dictionary of birds. London, Nelson. Tvolvi, J. 1980. Mammalian reproductive strategies: a generalized relation of litter size to body size. Oecologia 45: 39-44. VITT, L. J., g. A. $EIGEL. 1985. Life history traits of lizards and snakes. Amer. Natur. 125: 480-484. WESTERN, D. 1979. Size, life history and ecology in mammals. African J. Ecol. 17: 185-204. --, a: J. $$EMAKULA. 1982. Life history patterns in birds and mammals and their evolutionary interpretation. Oecologia 54: 281-290.