Body Size, Nest Predation, and Reproductive Patterns in Brown Thrashers and Other Mimids

Portland State University PDXScholar

Biology Faculty Publications and Presentations Biology

11-1986 Body Size, Nest Predation, and Reproductive Patterns in Brown Thrashers and Other Mimids

Michael T. Murphy Portland State University, [email protected]

Robert C. Fleischer

Follow this and additional works at: https://pdxscholar.library.pdx.edu/bio_fac

Part of the Biology Commons, and the Ornithology Commons Let us know how access to this document benefits ou.y

Citation Details Murphy, M. T., & Fleischer, R. C. (1986). Body size, nest predation, and reproductive patterns in Brown Thrashers and other mimids. Condor, 446-455.

This Article is brought to you for free and open access. It has been accepted for inclusion in Biology Faculty Publications and Presentations by an authorized administrator of PDXScholar. Please contact us if we can make this document more accessible: [email protected]. The Condor88:446-455 ? The CooperOrnithological Society 1986

BODY SIZE, NEST PREDATION, AND REPRODUCTIVEPATTERNS IN BROWN THRASHERS AND OTHER MIMIDS'

MICHAELT. MURPHY2 AND ROBERTC. FLEISCHER3 Departmentof Systematicsand Ecology,Museum of NaturalHistory, Universityof Kansas,Lawrence, KS 66045

Abstract. We describethe breedingbiology of BrownThrashers (Toxostoma rufum) in Kansas, and combine this with data from other temperate-zonebreeding Mimidae to characterizerepro- ductive patternsin this group.Brown Thrashers produced clutches of 3 to 6 eggs, but clutchesof 4 predominated.Most pairsraised 2 broodsper year.Incubation required between 13 and 14 days, and hatchingwas usually asynchronous.Though sample size was small, asynchronyappeared to increasein frequencytowards the end of the breedingseason. Nestlings grew rapidly,and in 10 days or less most pre-fledginggrowth was completed. Young fledgednormally at 11 days of age at 65% of aduli weight, but with the tarsi near adult size. Nestlings starved in 27% of nests, but predatorswere responsiblefor most nest failures.Overall nest success was 43%. BrownThrashers are typical of other temperate-zonemimids. Modal clutch sizes are of either 3 or 4 eggs and all species are multi-brooded.Mimids from the southwesternUnited States and Mexico lay normally 3 egg clutches, but elsewhere4 eggs are most common. Incubationlength and nestlinggrowth rate vary significantlywith adult weight,but on average,incubation is 3 days shorter and nestlings grow 36% faster than predicted. Relative incubation length and relative fledgingweight both declined significantlywith adult weight, whereas daily nest mortality rate increasedsignificantly with adult size. Althoughour data are consistent with the hypothesisthat heavy nest predationhas favored rapid nestlinggrowth and completion of development outside of the nest, rapid growth may also function in brood reduction.Present data are insufficientto exclude conclusivelyeither factor in the evolution of rapid developmentin mimids. Key words: Broodreduction; growth; hatching asynchrony; Mimidae; nest predation; Toxostoma.

INTRODUCTION either exploiting unpredictablefood supplies, Variability of growth rates and hatching pat- or sufferinghigh ratesof nest predation.Hatch- terns in altricial nestlings have been related ing asynchrony,however, possibly occurs for chiefly to featuresof their food supply and the other reasons(Richter 1982, Clarkand Wilson frequency of nest loss to predators. Growth 1985, Hussell 1985, Mead and Morton 1985). rates determine peak nestling energy demand In this report we describe the breeding bi- (O'Connor1977, Ricklefs 1984) and time spent ology of Brown Thrashers(Mimidae: Toxos- in the nest, thereby influencingboth the par- toma rufum)in eastern Kansas, including the ent'sability to eliminatestarving young through first data on nestling growth. Aspects of their brood reduction (O'Connor 1977), and the reproductivebiology have been documented probabilitythat predatorswill locate and de- in a portion of their range (Erwin 1935), but stroynests beforefledging (Lack 1968, Ricklefs only fragmentaryinformation exists for Brown 1969a, 1984). Hatching asynchronyresults in Thrashers breeding west of the Mississippi size differencesamong young which has tra- River (Gabrielson 1912, Johnston 1958). In ditionally been viewed as an adaptationto fa- conjunctionwith data on hatchingand growth cilitate brood reduction (Lack 1954, Ricklefs patterns, and sources of nestling mortality in 1965, Howe 1976, Richter 1984). It may also othertemperate-zone breeding mimids, we also shorten exposure time for nest contents, and describe and attempt to identify the selective give the earliest hatching young growth ad- basis for breedingpatterns in this group. vantages to increase their probabilities of es- One possible contributor to variability in capinga predationattempt on the nest (Hussell reproductionis body size (Ricklefs 1968, Rahn 1972, Clarkand Wilson 1981). Existingtheory et al. 1975, Blueweisset al. 1978, Westernand thus predicts the evolution of hatching asyn- Ssemakula 1982, Calder 1984). Comparative chrony and rapid nestling growth in species breedingstudies must thereforecontrol for differencesin size. Comparisonsof allometric(i.e., size-dependent) relations of specific taxa to Received22 November1985. Final acceptance 3 March "average,"empirically derived allometric re- 1986. lations are in fact preferableto single species 2 Departmentof Life Sciences,Indiana State University, comparisons because they are less subject to TerreHaute, IN 47809. error. Our results suggest that reproductive 3 Hawaiian EvolutionaryBiology Program,University of Hawaii, 3050 Maile Way, 310 Gilmore, Honolulu, HI patterns in mimids exhibit size dependence, 96822. but that a combination of ecological pressures

[446] MIMID REPRODUCTION 447 have probablyacted in concert to produce the A group of nests that survived incubation characteristicmimid pattern of rapid nestling was used to measure nestling growth. Most growth and short nest occupancy. nests were visited daily. Nestlings were iden- tified by clipping toenails at the first visit. At the firstand all subsequentvisits, nestlingswere weighedto the nearest0.1 g (50 or 100 g Pesola METHODS Scale) and tarsus and eighth primary lengths BROWN THRASHERS measured to the nearest 0.1 mm. Adult sizes were obtainedfrom specimensin the KUMNH Field studies were conducted from the end of from eastern Kansas. April through July, 1981 and 1982 in mod- erately grazed pasture located 6.5 km west of the city of Lawrence,Douglas County, Kansas INTERSPECIFICSTUDIES (38057'N and 95019'W).Scattered shrubs and We restrictedour analysisto species that breed trees were found throughoutthe site, but hab- in temperate-zoneregions. Our sample includ- itats with a closed canopy comprised less than ed all 10 species of Mimidaebreeding in North 5% of the total area. Virtually all nests were America,and one SouthAmerican species. Due located within an intensively studied area to varying degrees of completeness, sample measuringabout 740 x 540 m (40 ha). sizes for differentanalyses varied. We treated Nests were located by observing females in Arizonaand south Texaspopulations of Curve- transit to either existing nests or those under billedThrashers (T. curvirostre)separately since construction. We visited nests every 2 to 3 adult body size, clutch and egg sizes, and nest- days until eggs were laid, and then followed ling growth all showed distinct differences. them until fledgingof young or destructionof Adult weights were taken from original the nest. Dates of clutch initiation were ob- sourceswhen given. Otherwise,we used Durin's tained either by direct observationor by back- (1984) compilation, or the field records of as- datingfrom hatchingdate of clutches.Clutches sociates to obtain weightsfor adults. Adult tar- observed duringegg-laying were consideredto sus lengths were measured (nearest 0.1 mm) be complete if successive visits indicated no from 5 male and 5 female specimens for each change in egg number. Heavily incubated species with data on growthof the tarsus(study clutches were also assumed to be complete. skins from the KUMNH). We estimatedmean Because eggs were always laid on successive egg weight for each species using the egg mea- days during laying, and because we had no surements given in Bent (1948) and Fraga evidence for egg removal by the brood-para- (1985), and the conversion factors described sitic Brown-headedCowbird (Molothrus ater), above for Brown Thrashers.This was justified we assumed that clutchesfirst observed during by comparison of calculatedegg weight to ac- incubationrepresented fuli clutches.Nests that tual fresh egg weight for CrissalThrashers (T. fledged at least one nestling were considered dorsale)and Chalk-browedMockingbirds (M. successful. We corrected nest success for ex- saturninus)given by Finch (1982) and Fraga posure time using Mayfield's(1961) method. (1985), respectively.In both cases, calculated Additional clutch size data were obtained and observedweights differed by less than 1%. from nest recordsat the University of Kansas Clutch size, incubation and nestling period Museum of Natural History (KUMNH, Law- lengths, weight gain and tarsus growth, and rence, Kansas). For the nest recordsto be sat- nest successwere takenfrom originalliterature isfactory for use, we required that successive sources.We used Bent's (1948) summariesfor visits had been made to each nest that indi- the formerthree variables only when data were cated no changein egg number,or the observer not available from field studies. noted that incubation was in progress. Only Rates of nestling weight gain and increase 38 of 98 nest records satisfied these criteria. in tarsus length were calculatedfor all species We combined the nest record-cardinforma- with data using Ricklefs' (1967) graphical tion with our field data and groupednests into method. Allometric relationships between 15-day periods beginning 15 April to test for body size (or egg size) and reproductivetraits seasonal changes in clutch size. We also mea- in mimids were described by applying least suredeggs for maximumlength (L) and breadth squareslinear regression to double logarithmic (B) in six nests in the field in 1981 and for 21 transformations of each dependent variable clutches previously collected in the same re- versus body weight (or egg weight). Compar- gion and now located at the KUMNH. Volume ison of the mimid relationshipto established was calculated as 0.51(L x B2; Hoyt 1979), allometricbaselines (see below) were made by and then converted to weight by multiplying determining whether the 95% confidence in- by egg density (= 1.09; Manning 1978). tervals enclosing the mimid expression in- 448 MICHAELT. MURPHY ANDROBERT C. FLEISCHER

TABLE1. Summaryof the averageweight (g), and tarsus and eighthprimary lengths (mm) for nestlingBrown Thrashers in easternKansas. Reported values are the mean, with the standarddeviation and samplesize in parentheses.Hatching is day 1.

Age Weight Tarsus Eighth primary Day 1 5.5 (1.28; 19) 8.3 (0.84; 19) 2 9.2 (1.50; 18) 10.6 (0.92; 21) 3 13.0 (2.51; 19) 12.5 (1.73; 19) 0.5 (0.40; 15) 4 17.3 (2.84; 12) 15.0 (2.13; 11) 1.9 (0.44; 8) 5 24.8 (3.68; 21) 20.3 (1.75; 23) 4.5 (0.81; 9) 6 31.4 (2.85; 14) 24.0 (1.38; 19) 8.6 (1.10; 10) 7 36.6 (3.54; 14) 27.2 (1.17; 14) 13.6 (1.52; 10) 8 38.4 (4.33; 16) 29.1 (1.76; 16) 17.9 (1.10; 9) 9 44.1 (2.58; 15) 31.7 (2.05; 17) 22.3 (1.21; 10) 10 45.2 (3.45; 8) 32.9 (1.42; 8) 26.8 (1.14; 4) 11 45.7 (3.23; 3) 34.3 (0.59; 3) Adult 72.2 (5.19; 18) 35.0 (0.85; 18) cluded the baseline allometricrelationship for Mean egg mass was 5.8 g (SD = 0.55, n = birds. 27 clutches).Eggs measured in the field during 1981 were identical in size to from the RESULTS eggs KUMNH (x = 5.8 g, SD = 0.70, n = 6 nests BROWN THRASHERS and SD = 0.44, n = 21 nests, respectively). Egg-layingextended from the latter third of Using an adult weight of 72.2 g for Brown April until the end of July. Two of three fe- Thrashersin Kansas (SD = 5.19 g, n = 18), males that successfullyfledged their first broods and Rahn et al.'s (1975) prediction equation were known to have laid second clutches, and for passerines,predicted egg weight was 6.2 g. all females that failed in their firstattempt laid This exceeded the observed mean in Kansas replacementclutches. Most pairs probablyat- by about 7%, but the differencewas not sig- tempt to raise two broods per year (see also nificant (t = 0.71, df = 26, P > 0.40). Erwin 1935). The average length of incubation (time be- Clutch size averaged 4.1 eggs in both the tween the layingof the last eggand its hatching) field sample (SD = 0.70, n = 21, range = 3- was 13.6 days (SD = 0.55, 3 at 14 and 2 at 5) and the nest record cards (SD = 0.74, n = 13), significantlylonger than that reportedby 38, range = 3-6). Combiningdata sets, modal Erwin (x = 12.6 days, SD = 0.71, n = 17; t = clutch size in Kansas was 4 eggs (n = 33), 2.89, df = 15, P < 0.05). Though based on a followed in decreasingorder of abundanceby small sample size, it appeared that hatching clutches of 5 (n = 16), 3 (n = 9) and 6 (n = 1) synchrony varied seasonally. Of six nests for eggs. Smalleraverage clutch size was reported which the patternof hatchingwas determined, by Porter (in Johnston 1958) from northern two early-season nests were synchronous, Kansas (X = 3.5 eggs, n = 51). We speculate whereasfour clutcheshatched in the latterhalf that Porter included incomplete clutches, or of the season were all asynchronous(i.e., > 1 nests that had sufferedpartial losses. Ourover- day requiredfor all eggs to hatch). all mean for Kansas (4.1 eggs, SD = 0.68, n = The nestling period averaged 11.6 days 59) did not differ significantly(t = 1.55, df = (SD = 0.89 days, n = 5, range = 11-13), in 107, P > 0.20) from the mean clutch size of accordancewith Erwin's(1935) reportedmean Brown Thrashersin Tennessee (x = 3.9, SD = of 11 days. Table 1 summarizesdata on nest- 66, n = 50; Erwin 1935). ling weight gain and growth of the tarsus and Early and late clutches in Kansas averaged eighth primary feather over this period. Av- 4.0 eggs (SD = 0.72, n = 32; 15 April to 15 erageasymptotic weight (A) and rate of weight May, and 16 June throughJuly). Though mid- gain (K) were 47.9 g and 0.512, respectively. season clutches tended to be larger (x = 4.3 Average values for the tarsus were A = 34.9 eggs, SD = 0.63, n = 27; late May and early mm and K = 0.444. At fledgingthe tarsi were June), the difference between mid-season nearly adult length but weight was only at 65 clutches and those laid at other times was not to 70% of adult size (Table 1). The times re- significant(t = 1.87, df= 57, P = 0.07). Porter quired to grow between 10 and 90% of final (Johnston 1958) reportedthat clutch size also fledgingsize 90;Ricklefs 1967) were 8.6 and peaked duringthe same period. Erwin(1935), 9.9 days, respectively,(t1- for weight and tarsus however, reported a significant seasonal de- length.Hence, in 10 days or less both variables cline in clutch size in Tennessee, with no mid- were about fledgingsize, which for the tarsus season peak. was nearly adult size. MIMID REPRODUCTION 449

TABLE2. Summarystatistics for reproductivetraits in temperate-zonebreeding Mimidae. All weightsare in grams, and incubationlength in days. Eggweights were taken from Bent (1948) except for Chalk-browedMockingbirds. Egg weightswere also availablefor catbirds,Brown Thrashers and CrissalThrashers from field studies.These follow Bent's data, which is always the first listed under egg weight. Under source listings, KS, TX, and AZ refer to weights and breedingdata for populationsspecific to Kansas,Texas and Arizona.For clutch size and incubationlength, means are given outside of parentheses,followed by the standarddeviation and sample size in the parentheses.

Species Body weight Clutchsize Eggweight Incubationlength Source Dumetellacarolinensis 36.2 3.3 (0.75; 22) 3.97 - 1 3.8 (0.67; 73) 13.4 (-; 30) 2 3.9 (-; 37) 12.9 (-; -) 3 4.1 (-; 314) 3.95 - 4 Oreoscoptesmontanus 43.3 3.5 (0.80; 38) 4.47 15.0 (1.3; 9) 5 Mimus polyglottos 48.5 3.9 (0.61; 182) 4.52 12-12.5* 6 M. saturninus 78.7 3.6 (0.76; 14) 6.70 13.5 7 Toxostomarufum 68.8 3.9 (0.66; 50) 5.54 12.6 (0.71; 12) 8 72.2 4.1 (0.71; 59) 5.82 13.6 (0.55; 5) 9 (KS) T. longirostre 69.9 3.4 (0.80; 7) 5.95 14.0 (0.60; 3) 10 T. curvirostre 84.5 3.8 (0.50; 67) 5.85 14.0 (1.3; 18) 10 (TX) 79.4 3.0 (0.58; 56) 6.27 14 11 (AZ) 2.8 (0.56; 15) 12 (AZ) T. dorsale 62.7 3.0 (-; 14) 5.45, 5.5 14 13, 11 T. bendirei 62.2 usually 3 5.19 - 11 T. lecontii 61.9 3.1 (0.68; 22) 5.95 14-20 14, 11 T. redivivum 86.4 usually 3 7.52 14 11 * -Incubation periodunderestimated by 1 day (see text). Sources:1-Zimmerman 1963; 2-Nickell 1965; 3-Johnson and Best 1980; 4-Crowell and Rothstein 1981; 5-Reynolds 1981;6-Laskey 1962; 7- Fraga1985 and Mason 1985;8-Erwin 1935;9-this study;10-Fischer 1981; 11--Bent 1948; 12-Ricklefs 1965; 13-Finch 1982; 14-Sheppard 1970.

The two extreme examples of weight gain nificantly related (r = -0.340, n = 12, P >> are presented in Figure 1. Initial brood size 0.05), but regionalvariation in clutch size ap- was four young in both nests. In the early- pears to exist. Thrashersfrom the southwest- season nest, all nestlings hatched on the same ern U.S. and Mexico (Toxostoma plus Oreo- day, grew rapidly, approachedan asymptotic scoptes) lay fewer eggs per clutch (X = 3.12, weight by about day 10 and fledgedon day 11 SD = 0.263, n = 4) than mimids from else- (Fig. 1). Their averagerate of weight gain was where on the North American continent (X = K = 0.632 = 7.0 days). Nestlings in the 3.77, SD = 0.195, n = 5). The difference is second nest(to-90 hatched asynchronouslytowards significant(t = 4.2, df= 7, P < 0.01). Irrespec- the end of the breeding season. Weight gain tive of location, all species are double- or oc- was slower (K = 0.396, = 11.1 days), and casionally triple-brooded. by days 7 and 8 the thirdto-90 and fourth hatched Eggweight was a directfunction of body size. nestlings,respectively, died apparentlyof star- Adult body weight accounted for 87% of the vation. The fourth hatched nestling died de- total variation in egg weight (P < 0.001; Table spite being largerthan the remainingbirds (Fig. 3). The 95% confidence limits completely en- 1). Nestlings starved in 3 of 11 nests (27.3%), closed the average regression line describing all during the latter half of the nesting season. the relationshipbetween adult weight and egg Overall, 10.8%of nestlings starved, and in all weight in passerines(Fig. 2), indicatingno sig- cases they were the last hatched nestling(s)in nificant deviation from the average passerine each nest. relationship. Incubation length was less de- Of 24 nests, 43.7%fledged at least one nest- pendent on body size. In general, incubation ling (instantaneous mortality rate = 0.0291, length in birds correlates directly with adult with a 28-day nest period). Daily mortality body weight(Rahn and Ar 1974)and eggweight rate, M (Ricklefs 1969b), was 2.96%/day. Er- (Rahn et al. 1975, Western and Ssemakula win (1935) did not correctnest success for ex- 1982). The averagelength of incubationvaried posure time, but detected a similar probability only between 13 and 15 days in mimids (Table of success (37.5%).Nest predation accounted 2), and was the longest in the Sage Thrashers for 54%of nest failuresin Kansas, followed by (Oreoscoptesmontanus), despite it being the weather (27%)and desertion (9%).The cause second smallest species. Incubationlength was for four nest failures was undetermined. not significantly correlated with either adult = or = INTERSPECIFIC body weight (r 0.163) egg weight (r COMPARISONS 0.216). Body size and reproduction.All mimids have After excluding Sage Thrashers, however, modal clutch sizes of either 3 or 4 eggs (Table significant relationships were found to exist 2). Clutch size and body weight were not sig- between incubation length and both adult 450 MICHAELT. MURPHY ANDROBERT C. FLEISCHER

45 8 @13 -

7 .677 12 35 EW= .340BW Early Nest * o10 6-

25 5 -

-J 0 15 o LateNest

5 S.613 3

2 4 6 8 10 12 NESTLINGAGE (DAYS) FIGURE 1. The pattern of weight change of nestling 30 50 70 90 BrownThrashers in an early(closed circle)and a late nest ADULT WEIGHT (g) (open square)in Kansas in 1981. The d present on the growthcurves of two nestlingsfrom the late nest indicates FIGURE 2. A log-log plot of the relationshipbetween theirweights on the day priorto their disappearancefrom eggweight (EW) and adultbody weight (BW) in temperate- the nest. The late nest is displacedone day to the rightto zone breeding mimids (EWm). The solid, heavy line rep- facilitatepresentation. resentsthis relationship,whereas the dashedline indicates the "average"pattern between egg weight and body weight in passerinebirds (EWp;Rahn et al. 1975). The shaded region encloses the 95% confidence interval around the weight and egg weight (P < 0.05; Table 3). In mimid relationship.Numbers 1 through13, respectively, both cases, the exponentsdiffered significantly indicate Dumetella carolinensis,Oreoscoptes montanus, from zero (P < 0.05), but were also below the Mimus polyglottos,T. bendirei,T. dorsale, T. lecontii, T. exponents relatingincubation length to either rufum, T. rufum (Kansas), T. longirostre,T. curvirostre = (Texas), T. curvirostre(Arizona), M. saturninus,T. redi- adult weight (b 0.200, Rahn and Ar 1974) vivum. or egg weight (b = 0.217, Rahn et al. 1975; b = 0.200, Westernand Ssemakula 1982) in birds generally. In the former case, the difference The statistics of nestling growth are sum- between the exponent in mimids and other marized in Table 4. Our measure of growth birds was significant (P < 0.05). We cannot rate, tl,,90 (Ricklefs 1967), is an inverse mea- explain the relatively long incubation period sure of rate. Hence, large values indicate slow of Sage Thrashers,but suspect that environ- growth.Rate of weightgain varied significantly mental factors (e.g., low air temperature,as in with adultbody weight(r = 0.691), and asymp- Murphy 1983) lengthened the incubation pe- totic nestling weight (r = 0.691; P = 0.04, n = riod. 9). In neither of the equations relatinggrowth Rahn and Ar's (1974) equationprovided the rateto weight(Table 4) did the exponentsdiffer closest fit of observed to expected incubation significantly(P > 0.05) from Ricklefs' (1968) length of the several possible predictionequa- empirically derived exponent describing the tions (see above). The differencesbetween pre- relationshipbetween growth rate and asymp- dicted and observed length averaged2.9 days totic nestling weight in altricial birds (b = less than predicted (SD = 0.82, n = 9). The 0.278). deviation from the predictedincubation length However, the coefficient (Y-intercept) in was significantly correlated with adult body Ricklefs' equation was 36% higher than the weight (r = 0.909, Y = -7.7 + 5.91[log. values we obtained using either adult or WEIGHT], n = 9, P < 0.001). This was also asymptotic weights (Fig. 3). The 95% confi- true when Sage Thrasherswere excluded (r = dence limits around the regressionof rate of 0.917, Y = -5.9 + 4.99[logWEIGHT], P = weight gain on asymptotic nestling weight for 0.003). Hence, large species had relatively the mimids (Fig. 3) did not includeRicklefs' (1968) shortest incubation periods. allometricaverage for birds with altricialyoung, MIMID REPRODUCTION 451

TABLE3. Powerequations of the form Y = aXband statisticsdescribing variation in breedingtraits in the Mimidae. The dependentvariable (Y) in each regressionis given under Variable,followed by the independentvariable (X) in parenthesesas follows: BW = adult weight, EW = egg weight,AW = asymptoticnestling weight, WT = rate of weight gain, and TR = rate of tarsusgrowth. Weights are all in g except for adult body weightunder incubation length, which is in kg.

Variable n a b(95%CI) CI) r P Y(95% Eggweight (BW) 13 0.436 0.613 5.56 64.0 0.872 0.001 (0.456-0.770) (5.34-5.78) Incubation(EW) 8 11.06 0.121 13.62 5.59 0.563 0.035 (0.018-0.224) (13.35-13.89) Incubation(BW) 8 17.26 0.087 13.62 0.065 0.605 0.025 (0.019-0.154) (13.37-13.87) Weightgain (AW) 9 2.90 0.306 9.24 44.1 0.477 0.040 (0.027-0.585) (8.63-9.90) Weightgain (BW) 9 3.29 0.253 9.24 59.3 0.477 0.040 (0.022-0.484) (8.63-9.90) Tarsusgrowth (BW) 6 3.52 0.268 10.30 55.3 0.631 0.062 (0.005-0.531) (9.48-11.19) Nestling period (WT) 7 2,75 0.675 12.10 9.01 0.549 0.059 (0.006-1.344) (11.41-12.84) Nestling period (TR) 6 2.43 0.689 12.12 10.3 0.722 0.036 (0.139-1.239) (11.43-12.85) Columnheadings starting with n are:sample size, Y-intercept,regression coefficient (95% confidence limits below),mean of the dependentvariable (95% confidencelimits below),mean of the independentvariable, coefficient of determination,probability level.

indicatingthat nestlingmimids grewfaster than Rw appeared to vary inversely with adult expected based on size. weight. A test of association between Rwand Tarsusgrowth rate also scaled to adult body adult weightusing Kendall'scoefficient of rank weight with about the same exponent (b = correlation indicated that Rw varied signifi- 0.270; Table 3). The relationship only ap- cantlywith adult weight(tau = 0.61, P < 0.05). proached significance(r = 0.796, P = 0.062, Nestlings of large mimids (i.e., >60 g adult n = 6). The largercoefficient for tarsusgrowth weight) thus fledged relatively lighter than indicatesrelatively slower growth compared to nestlingsof the small species (< 50 g). This was weight, despite the fact that the young fledged not true of Rt, the ratio of asymptotic to adult with their tarsi at about adult size (Table 4). tarsuslength. Rt was at or near 1.0 in all species Presumably,this reflectsthe fact that the tarsi by the time of fledging(Table 4). were relatively much closer to adult size at Hatchingpatterns, starvation, and nest suc- hatching than was weight (e.g., Table 1). cess. Data on hatchingasynchrony and the oc- Relative weights at fledging (Rw) were all currenceof nestling starvationwere not avail- fairlylow (Table 4; compareto Ricklefs 1968). able for all species (Table 5). Of the 9 species

TABLE4. Growthrates (K, asymptoticsizes (A) and the ratios (R) of nestlingsize at fledgingto adult size for to-90), weight (g) and tarsus length (mm) in nestling mimids. Nestling period length is given in days. Values for weight gain are above and outside of parentheses,whereas those for tarsusgrowth are below and within the parentheses.

Nestling Adult size A K R Source Species tl0-90 period Dumetellacarolinensis 36.2 28.0 0.549 8.0 0.77 11 1 (24.0) (23.0) (0.468) (9.6) (0.96) 1 Oreoscoptesmontanus 40.1 34.1 0.543 8.0 0.85 12 2 (30.5) (33.2) (0.468) (9.4) (1.0) 2 Mimus polyglottos 47.7 39.1 0.452 9.7 0.82 12 3 50.0 37.5 0.492 8.9 0.75 4 (33.3) (32.2) (0.452) (9.7) (0.97)* 3 M. saturninus 78.7 61.6 0.476 9.2 0.78 12 5 Toxostomalongirostre 67.7 49.9 0.443 9.9 0.73 13 3 (36.5) (35.2) (0.408) (10.8) (0.96)* 3 T. rufum 72.2 47.9 0.512 8.6 0.66 11 6 (35.0) (34.9) (0.444) (9.9) (1.0) 6 T. curvirostre 79.7 55.0 0.384 11.5 0.69 - 7 84.5 55.6 0.444 9.9 0.65 14 3 (34.5) (33.5) (0.344) (12.8) (0.97)* 3 * Fischer(1983) obtainedlower valuesfor R, presumablybecause of slightdifferences in our techniquefor measuringtarsus length. Sources:1--Zimmerman 1963; 2-Killpack 1970;3-Fischer 1983;4-Breitwisch et al. 1984;5-Fraga 1985;6-this study;7-Ricklefs 1965. 452 MICHAELT. MURPHY AND ROBERTC. FLEISCHER

The percentageof nests which successfully fledged young varied from 26% to 70%. To compare species for differencesin the rate of .21278 we calculatedthe 3.94Weight278 nest loss averageproportion Ratea= (unweighted) of successful nests for each species, and converted it to averagedaily nest 10 mortality rate (Ricklefs 1969b). A regression of averagedaily nest mortalityrate (NMR) on the logarithmof adult weight yielded a signif- 5 icant positive relationship (r = 0.727, n = 8, = a P 0.04; NMR = -0.069 + 0.0559[log- 1 I- WEIGHT]. To our knowledge, these are the firstdata for any passerinefamily demonstrat- .306 = ing a correlationbetween body size and repro- Ratem 2.90Weight ductive success.

30 40 50 60 DISCUSSION ASYMPTOTICWEIGHT (g) Breedingpatterns of Brown Thrashersin Ten- nessee and Kansaswere very similar.The only FIGURE3. The rateof nestlingweight gain (to9; Rick- apparent differences between Erwin's lefs 1967) plotted against asymptotic nestling weight of (1935) temperate-zonebreeding mimids. The lower, heavy line study and ours were that incubation was sig- describesthe relationshipin mimids (Ratem),whereas the nificantlyshorter in Tennessee,and that clutch upperline representsthe "average"pattern for birds with size declined seasonally in Tennessee but not altricialyoung (Ratea;Ricklefs 1968). The shaded region in Kansas. The differencein incubationlength representsthe 95% confidenceregion around the mimid seemed to have arisen from method- relationship.Numbers 1 through9, respectively,indicate likely Dumetella carolinensis, Oreoscoptesmontanus, Mimus ological differences in the determination of polyglottos(from Texas), M. polyglottos(from Florida), length. Whereaswe counted incubation from Toxostomarufum, T. longirostre,T. curvirostre(Arizona), the end of egg-laying to hatching of the last T. curvirostre M. saturninus. (Texas), egg, Erwin probably counted to the hatching of the firstegg. The differencein seasonalvari- forwhich informationon hatchingpatterns was ation in clutch size was probably artifactual found, 7 reportedasynchrony to be common, since a definite trend existed for late clutches especiallyin largeclutches. Starvationwas not to be smaller in Kansas. Small sample size as well documented.For the seven specieswith rather than intrinsic population differences data, starvation occurredfrequently in three, probablyaccount for this discrepancy. was absent in two others, and in both the Gray Breeding patterns among temperate-zone Catbird (Dumetella carolinensis) and Curve- mimids were also quite uniform. Our empha- billed Thrasher was reported to be common sis on relatingreproductive traits to body size in one study yet essentially absent in another clearly showed that much of the observed in- (Table5). When starvationwas reported,about terspecificvariability in certain traits was due 25% of all nests were affected(Table 5). to differences in body size. However, our

TABLE5. Hatchingpatterns, observations on nestlingstarvation, and nest successfor mimids breedingin temperate- zone regions.Asyn and Syn referto asynchronousand synchronoushatching, respectively. Percentages under starvation refer to the percentageof nests with nestlingsthat experiencedstarvation. Under nest success, the numbersindicate the percentageof nests that fledgedat least one nestling.Numbers in parenthesesfollowing data indicate sources.

Species Hatchpattern Nestlingstarvation Nest success Dumetellacarolinensis Asyn (1, 2) Very low (3) 58 (1), 70 (4), Present(15.1%; 4)* 69 (5), 44 (6) Mimus polyglottos Unknown Absent (7) 56 (7) M. saturninus Asyn (8) Present(20.0%; 8) 27 (8) Oreoscoptesmontanus Syn (9) Present(- 30%;9) 45 (9) Toxostomarufum Asyn (10) Present(27.3%; 10) 44 (10), 37 (11) T. longirostre Syn (7) Absent (7) 26 (12) T. curvirostre Asyn (7) Absent (7) 37 (12) Asyn (13) Present(?; 13) Unknown T. dorsale Asyn (2, 14) Unknown 48 (14) T. lecontii Asyn (15) Unknown Unknown * The figureof 15.1%refers to percentageof nestlingsthat starved.A highervalue wouldoccur if expressedin relationto numberof nests. Sources:1--Johnson and Best 1980;2-Bent 1948;3-Johnson and Best 1982;4-Kendeigh 1942;5-Slack 1976;6-Best and Stauffer1980; 7-Fischer 1983;8-Fraga 1985;9--Reynolds 1981; 10--this study;11-Erwin 1935; 12-Fischer 1981; 13-Ricklefs 1965; 14--Finch 1982; 15-Sheppard 1970. MIMID REPRODUCTION 453

method of analysis also highlighteda number predation in mimids favors short nest occu- of unambiguoustrends in the data indicating pancy and completion of growthoutside of the that ecological factors have significantlyinflu- nest. Emphasis on leg growth and the attain- enced the evolution of life histories in the ment of functionalmaturity of the legs of nest- Mimidae. lings ensuresthat early fledgingis possible, es- Clutch size, for example, was independent pecially in ground-foraging birds. Chalk- of body size yet did vary with geographicdis- browed Mockingbird nestlings, for example, tribution. All species were multibrooded, and can successfullyfledge at 9 days of age if dis- laid normally 3 to 4 eggs per clutch. Mimids turbed by predators, though nest departure of the arid southwestern U.S. and Mexico, normallyoccurs 3 to 5 days later (Fraga1985). however, produced smaller clutches than Dark-eyed Juncos (Junco hyemalis) also ex- species breeding elsewhere. This was evident perienceheavy nest predation,and exhibit very even within Curve-billedThrashers (Table 2). similar growth patterns (Smith and Andersen Nestling Curve-billedThrashers from Arizona 1982; see also Austin and Ricklefs 1977). also showed the slowest relative growth rate We suspect,however, that otherfactors have of all species in the sample (Fig. 3). The small also contributed to the evolution of nestling average clutch size of these species, and the mimid growthpatterns. For example, hatching heavy nestling starvation in Curve-billed asynchrony is typical of most mimids, and Thrashersfrom Arizona (Ricklefs 1965), but along with rapidnestling growth, comprise the not Texas (Fischer 1983), suggestthat the rate essential components of the brood reduction at which offspringcan be supplied with food strategy(O'Connor 1977). This paradigmpro- is limited in desert environments. This may poses that when parents are faced with poor be due either to low habitat productivity (Ro- food suppliesthat are temporallystable, starv- senzweig 1968) or limitation of adult activity ing nestlings should be eliminated rapidly to by thermalstresses (e.g., Calder 1968, Ricklefs avoid feeding young that are destined to die and Hainsworth 1968, Austin 1976, 1978). (O'Connor 1977). Hatching asynchrony pro- The strongestpattern, however, was the con- duces size differencesthat are quickly accen- sistent trend for rapid development of both tuated by rapid growth. In this regard,rapid embryos and nestlings,and the generallyshort growth is critical because it increases nestling periods of nest occupancy. Incubation length energy demands (Ricklefs 1984) and maxi- did not scale closely to body size in our sample, mizes intrabrood competition, which even- due mainly to the long incubation period of tually kills the starving offspring.Patterns of Sage Thrashers. In all species, however, in- nestling starvation that conform to the brood cubation was shorter than expected based on reductionmodel have been observedin Curve- either adult size or egg weight. Relative incu- billedThrashers (Ricklefs 1965), Chalk-browed bation length also decreased significantly as Mockingbirds (Fraga 1985) and Brown body size increased, hence, the large species Thrashers(this study).Gabrielson's (1912) ob- had the shortestrelative lengths of incubation. servations on the distribution of food to four Nestling growth rate was dependent on body nestling Brown Thrashersalso match predic- size, but nestlings nonetheless grew signifi- tions of the brood reductionmodel in that the cantly faster than expected. Growth patterns smallest nestling received significantly fewer were such that young fledged at only 60% to feeds than expected by chance (n = 878 feed- 80% of adult weight, but with the tarsi always ings, G = 11.7, df= 3, P < 0.01). near adult size. Relative weight at fledgingalso Thus, rapid nestling growth (and hatching decreased as adult body size increased, indi- asynchrony)in BrownThrashers, and possibly catingthat the young of the largespecies fledged other mimids, may also function in brood re- at relativelyearlier stages of development than duction. Though nestling starvation is less the offspringof small species. common than nest predation among mimids, These data, and the finding that daily nest the variableoccurrence of starvationmay itself mortality rate rose significantlyas adult body reflectthe temporallyand/or spatiallyvariable size increased, suggest strongly that minimi- natureof food supplies.At present,we cannot zation of the time spent in the nest exposed to identify which, if either, of these models (nest predatorsas vulnerableeggs or nestlingsis ex- predationvs. brood reduction)is primarilyre- tremely important for mimid reproductive sponsible for the patternof rapid embryo and success. In accordance,predation was the ma- nestlinggrowth that we have detected. We feel jor cause of nest failure in all the studies cited it is likely that both have contributed to the in Table 5 (see also Biedenweig 1983), except evolution of nestling mimid growth patterns. for Arizona Curve-billed Thashers (Ricklefs This conclusion, though unsatisfyingfrom the 1965). Our findingsare thereforein agreement standpointof clean hypothesistesting, perhaps with Fischer's(1983) proposalthat heavy nest more realisticallyreflects the multiple selective 454 MICHAELT. MURPHY ANDROBERT C. FLEISCHER pressures impinging on individual reproduc- BEST,L. B., ANDD. F. STAUFFER.1980. Factorsaffecting tive success. We suspect that many ground- nestingsuccess in riparianbird communities.Condor birds nest close to the 82:149-158. foraging (which ground BIEDENWEG,D. W. 1983. Time and energy budgets of and frequently experience heavy nest preda- the Mockingbird (Mimus polyglottos) during the tion), face similar selection pressures,and ex- breedingseason. Auk 100:149-160. hibit patterns similar to those found in the BLUEWEISS,L., H. Fox, V. KUDZMA,D. NAKASHIMA, R. Mimidae. PETERS,AND S. SAMS. 1978. Relationshipsbetween tests of body size and some life historyparameters. Oecologia Clearly,more data and experimental (Berl.)37:257-272. these ideas are necessary.Information is need- BREITWISCH,R., P. G. MERRITT,AND G. H. WHITESIDES. ed on hatching patterns, nestling growth, and 1984. Why do NorthernMockingbirds feed fruit to rates of starvationin all species, but especially their nestlings?Condor 86:281-287. for mimids of the desertsof southwesternNorth CALDER, W. A.1968. The diurnalactivity of the road- runner,Geococcyx californianus. Condor 70:84-85. America. It would be instructive also to de- CALDER,WM. A., III. 1984. Size, function and life his- termine whether food supplies, hatching pat- tory. HarvardUniv. Press, Cambridge,MA. terns, the frequencyof nestling starvation,and CLARK,A. B., ANDD. S. WILSON. 1981. Avian breeding reduction predatoractivity covary on either an annual adaptations:hatching asynchrony,brood this and nest failure.Q. Rev. Biol. 56:253-277. or seasonalbasis (Meadand Morton 1985, CLARK,A. B., AND D. S. WILSON. 1985. The onset of study). Such information may allow discrim- incubationin birds. Am. Nat. 125:603-611. ination between the several hypotheses that CROWELL,K. L., ANDS. I. ROTHSTEIN.1981. Clutchsizes have been proposed to account for the evo- and breedingstrategies among Bermudanand North lution of hatching asynchrony and nestling Americanpasserines. Ibis 123:42-50. DUNN, J. B., III. 1984. Body weights of 686 species of growth rates in passerinebirds. North American birds. Western Bird-Banding, Monogr.No. 1. SUMMARY ERWIN,W. G. 1935. Some nesting habits of the Brown Sci. 10:179-204. of in Brown Thrasher.J. Tenn. Acad. Our analysis reproduction FINCH,D. H. 1982. Rejectionof cowbirdeggs by Crissal Thrashersand other temperate-zonebreeding Thrashers.Auk 99:719-724. mimids leads us to conclude that (1) body size FISCHER,D. H. 1981. Breedingbiology of Curve-billed accounts for much of the interspecific vari- Thrashers and Long-billed Thrashers in southern in certain but that Texas. Condor 82:392-397. ability reproductivetraits, FISCHER,D. H. 1983. Growth, development, and food (2) ecological pressureshave selected for dis- habitsof nestlingmimids in southTexas. Wilson Bull. tinct patterns,including rapid development of 95:97-105. embryos and nestlings. We suggest that both FRAGA,R. M. 1985. Host-parasiteinteractions between loss of nests to and unpre- Chalk-browedMockingbirds and shiny cowbirds.Or- frequent predators, nithol. Monogr. No. 36, American Ornithologists' dictable (but stable) food supplies have fa- Union, Washington,DC. vored the evolution of the rapid develop- GABRIELSON,I. N. 1912. A study of the home life of the mental rates that characterizetemperate-zone Brown Thrasher,Toxostoma rufum (Linn.). Wilson Mimidae. Hatching asynchrony is also com- Bull. 24:65-94. HOWE,H. F. 1976. Egg-size,hatching asynchrony, sex mon in mimids, and appears to function in and brood reductionin the Common Grackle.Ecol- broodreduction. Further data, and tests of these ogy 57:1195-1207. hypothesesare necessary,particularly in desert HOYT,D. F. 1979. Practicalmethods of estimatingvol- species. ume and fresh weight of bird eggs. Auk 96:73-77. HUSSELL, D.J.T. 1972. Factors affectingclutch size in arctic passerines.Ecol. Monogr.42:317-364. ACKNOWLEDGMENTS HUSSELL,D.J.T. 1985. Optimalhatching asynchrony in on Richter's of Clarkand We are to LouisBest, RichardJohnston, Jay Shep- birds: comments critique grateful Wilson'smodel. Am. Nat. 126:123-128. pard, and Peter Stacey for commentingon differentver- AND L. BEST. 1980. sions of this Louis Best in particularpointed JOHNSON,E. J., B. Breedingbiology manuscript. of the Catbirdin Iowa. Iowa State J. Res. 55: out the previousearly work on BrownThrashers. John L. Gray Zimmermanand an reviewersharpened our 171-183. anonymous E. ANDL. B. BEST. 1982. Factors interpretationof the data. We are appreciativefor all of JOHNSON, J., affecting their contributions. feedingand broodingof Gray Catbirdnestlings. Auk 99:148-156. JOHNSTON,R. F. 1958. Breedingof the BrownThrasher LITERATURECITED in Kansas.Bull. Kansas Ornithol.Soc. 9:17-18. AUSTIN,G. T. 1976. Behavioraladaptations of the Ver- KENDEIGH,S. C. 1942. Analysis of losses in the nesting din to the desert. Auk 93:245-262. of birds. J. Wildl. Manage.6:19-26. AUSTIN,G. T. 1978. Daily time budgetof the postnesting KILLPACK,M. L. 1970. Notes on SageThrasher nestlings Verdin.Auk 95:247-251. in Colorado.Condor 73:486-488. AUSTIN,G. T., ANDR. E. RICKLEFS.1977. Growth and LACK,D. 1954. The naturalregulation of animal num- developmentof the Rufous-wingedSparrow (Aimo- bers. OxfordUniv. Press, London. phila carpalis).Condor 79:37-50. LACK,D. 1968. Ecologicaladaptations for breedingin BENT,A. C. 1948. Life historiesof North Americannut- birds. Methuen,London. hatches, wrens, thrashersand their allies. U.S. Natl. LASKEY,A. R. 1962. Breedingbiology of Mockingbirds. Mus. Bull. 195. Auk 79:596-600. MIMID REPRODUCTION 455

MANNING, T. H. 1978. Measurementsand weights of RICKLEFS,R. E. 1965. Brood reduction in the Curve- eggs of the CanadaGoose, Branta canadensis,ana- billed Thrasher.Condor 67:505-510. lyzed and comparedwith those of other species. Can. RICKLEFS,R. E. 1967. A graphicalmethod of fittingequa- J. Zool. 56:676-687. tions to growthcurves. Ecology48:978-983. MASON,P. 1985. The nestingbiology of some passerines RICKLEFS,R. E. 1968. Patternsof growth in birds. Ibis of Buenos Aires, Argentina.Ornithol. Monogr. No. 110:419-451. 36, AmericanOrnithologists' Union, Washington,DC. RICKLEFS,R. E. 1969a. Preliminarymodels for growth MAYFIELD,H. 1961. Nesting successcalculated from ex- rates of altricialbirds. Ecology 50:1031-1039. posure. Wilson Bull. 73:255-261. RICKLEFS,R. E. 1969b. Patternsof mortalityin the nest- MEAD,P. S., ANDM. L. MORTON.1985. Hatchingasyn- ing of birds. Smithson. Contrib.Zool. 9:1-48. chronyin the MountainWhite-crowned Sparrow (Zo- RICKLEFS,R. E. 1984. The optimizationof growthrate notrichialeucophrys): a selected or incidental trait? in altricialbirds. Ecology 65:1602-1616. Auk 102:781-792. RICKLEFS,R. E., ANDF. R. HAINSWORTH.1968. Tem- MURPHY,M. T. 1983. Ecologicalaspects of the repro- perature dependent behavior of the Cactus Wren. ductive biology of Eastern Kingbirds: geographic Ecology49:227-233. comparisons.Ecology 64:914-928. ROSENZWEIG,M. L. 1968. Net primaryproductivity of NICKELL,W. P. 1965. Habits, territory,and nesting of terrestrialcommunities: predictions from climatolog- the catbird.Am. Midl. Nat. 73:433-478. ical data. Am. Nat. 102:67-74. O'CONNOR, R. J. 1977. Growth strategiesin nestling SHEPPARD, J. M. 1970. A study of LeConte'sThrasher. passerines.Living Bird 16:209-238. Calif. Birds 1:85-94. RAHN, H., AND A. AR. 1974. Theavian egg: incubation SLACK,R. D. 1976. Nest guardingbehavior by male Gray time and water loss. Condor 76:147-152. Catbirds.Auk 93:292-300. RAHN,H., C. V. PAGNELLI,AND A. AR. 1975. Relation SMITH, K. G., AND D. C. ANDERSEN.1982. Food, pre- to avian egg weightto body weight.Auk 92:750-765. dation and reproductiveecology of the Dark-eyed REYNOLDS, T. D. 1981. Nesting of the Sage Thrasher, Junco in northernUtah. Auk 99:650-661. Sage Sparrowand Brewer'sSparrow in southeastern WESTERN, D., AND J. SSEMAKULA. 1982. Life historypat- Idaho. Condor 83:61-64. terns in birds and mammals and their evolutionary RICHTER,W. 1982. Hatchingasynchrony: the nest-fail- interpretations.Oecologia (Berl.) 54:281-290. ure hypothesis and brood reduction.Am. Nat. 120: ZIMMERMAN, J. L. 1963. A nesting study of the catbird 828-832. in southernMichigan. Jack-Pine Warbler 41:142-160. RICHTER,W. 1984. Nestling survival and growthin the Yellow-headed Blackbird, Xanthocephalusxantho- cephalus.Ecology 65:597-608.

IS A LIFEMEMBERSHIP A GOODINVESTMENT ? TheCooper Society recently reinstated the Life Membership category. Your$600, paidin 4 installmentsor one lumpsum, adds to ouren- dowmentmanaged by the InvestingTrustees. The income generated willsupport publication of TheCondor. Your life membershipis an ex- cellentinvestment in andfor theSociety: it increasesvalue of the En- dowmentFund, thus defraying costs thatotherwise must be covered byannual dues. If eachmember had to payhis or her share of actually runningthe Society, dues would be morethan $40 peryear. By investing ina LifeMembership, you facilitate a broadmembership, the continued vitalityof the Society,and world-widecirculation of our scientific publication. Sinceyou are also concerned with your personal economics, consider thatwith the present$ 18 annualdues, you would break even with a lifemembership in34 years.Of course, dues will increase over the years, greatlyshortening the break-even period. For most ornithologists, a life membershipwill protect you against future dues increases, and during yourlifetime will be a bargain. Giftsand bequests to the Societyalso are invested.If you wishto makea gift or to mentionthe Society in yourwill, feel freeto contact any Officer,or C.John Ralph, Endowment Fund-Raising Committee, 7000 LanphereRoad, Arcata, California 95521 (707) 822-3691 or 822-2015.