The Annual Cycle of the American Goldfinch

Condor, 80:401406 0 The Cooper Ornithological Society 1978

THE ANNUAL CYCLE OF THE AMERICAN GOLDFINCH

ALEX L. A. MIDDLETON

The American Goldfinch ( Cmduelis t&is) was used to determine hormonal condition (Marshall is noteworthy for its unusually late breeding and Coombs 1957) and the phases of the testicular cycle (Marshall 1959). Due to the small sample sizes season (Tyler 1968, Immelman 1971, Middle- obtained from any collection period, gonadal data were ton 1972, Mundinger 1972, Newton 1972, von pooled in the analysis of the annual gonadal cycle. Haartman 1972)) and for its prenuptial (pre- In addition to birds collected for gonadal study, alternate) body molt, which is unique among molt data were also obtained from an additional 3,433 cardueline finches (Newton 1972, Middleton individuals banded at Guelnh between mid-November an d mid-May, 1970 to 1975, and from 371 birds 1977). Although late nesting was at one time trapped at nests during July and August, 1968 to thought to result from the need for thistle down 1975. Details of molt are given in Middleton ( 1977). (Forbush and May 1939), it was more re- cently suggested that some environmental RESULTS factor such as photoperiod may regulate the THE GONADAL CYCLE timing of gonadal development, the flowering and seeding of certain Compositae serving as Testicular cycle. Unwittingly I carried out my the ultimate factor which stimulates nesting study at approximately the same time as from year to year (Holcomb 1969, Newton Mundinger (1972). With a few notable dif- 1972). No one has thus far suggested that ferences, the size (Fig. 1) and histology of late gonadal development may be regulated the testes were similar to those in New York by some other annual event. State (Mundinger 1972). However, I give the In this paper, details of the major events of results of my study here as they can be related the annual cycle of the American Goldfinch directly to the other events of the annual cycle at Guelph, Ontario, are outlined. The sched- at Guelph. uling of these events, within the annual cycle, At Guelph old males (birds in at least their suggests that the prenuptial molt may serve second breeding season) come into breeding as the main physiological factor which pre- condition (stage 6 testes, Scott and Middleton cludes early breeding by this species. 1968) approximately two weeks in advance of first-year birds; testes of first-year birds were METHODS AND MATERIALS generally lighter, though nonsignificantly so (P > 0.05; students ’ t-test), than those of old Between May 1967, and November 1970, 118 nlale birds (Table 1) ; and old males lost breeding and 82 female goldfinches were shot within an 8 km condition (testes at stage 8) slightly in ad- radius of Guelph, for gonadal examination. At collection each bird was sexed, and the males were vance of first-year birds. aged ( Middleton 1974a). From June until September, The seasonal development of seminal sacs birds were collected at bi-weekly intervals, and at was reflected in development of the cloaca1 monthly intervals between September and May. protuberance. Basically, the weight cycle of Immediately after death, the left testis and seminal sac, or ovaries an d oviducts were dissected and seminal sacs was similar to that of testes, but placed in Allens’ fluid for 24 h. The right testis peak weight of the former was reached slightly and right seminal sac were placed in 10% formalin. later than in testes (Table 2). The histological After fixation for 24 h all excess tissues were re- cycle of seminal sacs has been described else- moved, the organs blotted dry and weighed to the where (Middleton 1972). nearest 0.1 mg on an Oertling R-20 balance. Wher- ever possible, diameters of the largest ovarian follicles The ovarian cycle. No attempt was made to were measured to the nearest 0.1 mm with a vernier study detailed histology of ovaries, and the caliper. Materials fixed in Allens’ fluid were embedded in wax and sections cut at 6~. Ovaries in ovarian cycle was assessed on follicular size which the follicular diameters could not be mea- (Fig. 2). Based upon a follicular diameter sured macroscopically were serially sectioned to per- of > 1.0 mm (Middleton 1971), breeding con- mit microscopic measurement of the largest follicle. dition was reached by the population in mid- Materials fixed in formalin were embedded in gelatin and frozen sections were cut at 10~. Sections were July and lasted until mid-August. However, mounted and either stained with Weigerts’ haematoxy- first follicles of > 1.0 mm were collected from lin and eosin (Allens’ fixed material) or with lipo- individuals on 6 July 1967, 28 June 1968, and uhvllic Sudan IV and Weigerts’ haematoxvlin (For- I ~ 12 May 1969, respectively. Maximum ovarian &in fixed materials). - development occurred in the latter half of July The histological condition of left testes was determined using techniques of Scott and Middleton when all the female birds were in breeding (1968) while lipid distribution within right testes condition. Thereafter, ovarian activity waned

[4011 402 ALEX L. A. MIDDLETON

6 7

I_ J FMAMJ JASOND JFMAMJJASOND MONTH MONTH FIGURE 2. Ovarian cycle of the American Gold- FIGURE 1. Testicular cycle of the American Gold- finch at Guelph, Ontario, based on size of largest fol- finch at Guelph, Ontario, based on weight of left licle. See Figure 1 for interpretation of symbols. testis. Vertical bars = range; horizontal bars = 2 SE; numbers give sample size. duct weight dropped markedly and regressed rapidly in September with the termination of and by September no females had ovaries in breeding activity (Table 2). breeding condition (Fig. 2). The seasonal development of oviducts MOLT paralleled that of ovaries (Table 2). Oviducts were minute between October and May, en- Although the molt cycle of the goldfinch has larged slowly during June, then rapidly been described elsewhere (Middleton 1977) reached maximum development in the second certain aspects emerge as being important half of July at the height of the egg-laying to this discussion. Both postnuptial and pre- period ( Middleton 1978). During August ovi- nuptial molts are prolonged, and it is difficult

TABLE 1. Weights of left testes of old and first-year goldfinches collected May-August, at Guelph, Ontario, 1967-70.

Weight (mg 1 cpx;Ig Age N Mean C SE Range

*May Old 5 7.8 k 5.2 0.7 - 28.5 lst-year 11 4.9 -I- 4.4 0.3 - 14.9

June old 9 25.9 k 10.3 0.3 - 78.5 1-15 lst-year 6 21.0 f 9.6 0.3- 57.4

June Old 7 85.3 k 9.7 41.5 - 115.2 16-30 lst-year 13 71.7 k 7.8 38.7 - 123.2

July Old 5 96.9 2 8.5 71.2 - 119.7 1-15 lst-year 10 99.3 k 9.2 64.2 - 143.4

July Old 2 114.3 k 24.1 90.2 - 138.4 16-31 lst-year 2 108.4 k 0.6 107.8 - 108.9

*August Old 3 115.7 2 9.0 98.1- 128.0 1-15 lst-year 3 104.4 2 8.9 93.7 - 122.0

August Old 2 110.2 2 11.0 99.1- 121.3 16-31 lst-year 2 58.8 k 50.5 8.3 - 109.3

* Discrepancy in sample size with Figure 1 arose through loss of age identity of samples. ANNUAL CYCLE OF THE AMERICAN GOLDFINCH 403

to determine exactly when one ends and the c-4-d 50% other begins (Fig. 3). However, there are - 51 75% -7c 100% peaks of molting activity which are similar to POSTJUVENAL those of the other temperate zone species C-W--- (Middleton 1977). Also, in the prenuptial __~~.___ __PRENUPTlAL MOLT molt, body molt begins with marked regularity in the second week of March (Middleton 1977). Molt ends abruptly at about the time

nesting starts. The nesting season is the only MONTH time of the year when no molt was found within the goldfinch population (Fig. 3). FIGURE 3. Chronology of the major events of the annual cycle of the American Goldfinch at Guelph, Ontario. MIGRATION Although Wiseman (1975) found no evidence Mundinger ( 1972)) my study shows that for migration by goldfinches at Cincinnati, male goldfinches apparently reach breed- Ohio, banding results at Guelph indicate that ing condition about two weeks before those the breeding population leaves the area dur- in New York. However, this may have been ing winter and is replaced by a distinct win- a function of Mundingers’ small samples and tering population, In addition, recovery in irregular collecting schedule. At Guelph, old Louisiana of a goldfinch banded in Guelph, males reach breeding condition before first- and the recovery records of the North Amer- year males and apparently lose breeding con- ican Bird Banding Scheme (courtesy of Ca- dition slightly before first-year birds. Al- nadian Wildlife Service) prove that some though Mundinger separated birds according goldfinches do undertake lengthy migrations. to age he did not comment on the effect of age The disappearance and reappearance of color- on the testicular cycle. banded summer birds, coincident with the re- The details of the testicular cycle fit nest- appearance and disappearance of winter ing data at Guelph which showed that old banded individuals, indicate that migration of birds (those in at least their second breeding the Guelph goldfinch populations occurs be- season) tend to build the earliest nest and tween early May to mid-June and between produce the first clutches, while presumed late October to mid-December (Fig. 3). first-year birds build most of the late August nests (Middleton in press). Also, the testes of DISCUSSION old birds were generally heavier than those In addition to corroborating the histophysio- of first-year birds as with some other avian logical details of the testicular cycle given by species (Scott and Middleton 1968, Payne

TABLE 2. Weights (mg) of left seminal sacs, and oviducts from goldfinches collected May-August, at Guelph, Ontario, 1967-70.*

Seminal sacs Oviducts

Month N Mean 2 SE Range N Mean f SE Range

May - - - 4 3.8 -c 0.8 1.7- 5.3

June 2 5.1 k 1.9 3.2 - 6.9 7 7.6 & 1.6 2.8 - 15.1 1-15

June 12 8.4 ” 1.8 1.8 - 16.7 17 15.8 4 2.7 4.4 - 46.9 16-30

July 9 13.5 c 2.2 8.2 - 25.6 10 154.0 -t- 63.5 10.6 - 465.6 1-15

July 5 19.6 k 2.4 15.0 - 28.6 8 186.9 r 117.6 1.9 - 949.3 16-31

August 4 24.0 2 2.6 18.7 - 30.8 4 45.9 -i- 15.3 13.9 - 87.6 1-15

August 3 26.4 C 7.1 13.1 - 37.1 - - 16-31

September 4 3.0 2 0.5 1.7- 4.2 6 12.0 -c 1.4 8.6 - 17.0

* Age classes not considered in these analyses. 404 ALEX L. A. MIDDLETON

1969, Erskine 1971). Although testicular size However, as molt of the flight feathers is cannot be equated precisely with gonadal completed in the goldfinch before migration condition (Middleton 1971, Mundinger 1972) occurs, capability for flight is not impaired. it may have a bearing on the apparent high Because different food constituents are needed fertility of old males (Middleton in press). for molt and deposition of fat, there is little Because the seminal sacs function as storage overlap in the demands of molt and migratory organs (Middleton 1972, 1974b) it was not preparation (Evans 1969). Therefore, the surprising that their development should fol- overlap of molt and migration probably does low the testes. The attainment of peak weight not impose an undue energetic strain on the in late August was probably due to the ac- goldfinch. cumulation of fluids and debris concomitant Also noteworthy in the annual cycle are the with the onset of testicular regeneration (Mid- markedly regular timing of sperm production, dleton 197413). prenuptial body molt (Middleton 1977), and The ovarian cycle closely follows the testic- migration from year to year (Middleton, ular cycle, though the majority of males reach unpubl. data). Little doubt remains that sea- breeding condition earlier than the females, as sonal variations of physiological and be- in most avian species (Lofts and Murton havioral phenomena are partly controlled by 1973). However, females vary more than endogenous rhythms in various animal groups males in the time at which they first reach (Gwinner 1973, 1975). In birds, the cycles breeding condition, because the date on which of reproduction, molt, and migration depend the first ovarian follicles of > 1.0 mm diam- upon such circannual rhythms (Gwinner eter were collected varied by almost two 1973). However, if such circannual rhythms months. Such variability may help to explain are deprived of the appropriate environmental the reports of abnormally early nesting by periodicities, they show periods which differ goldfinches (Nice 1939, Trautman 1940, from the environmental cycles which they Stokes 1950, Sutton 1959). normally match (Payne 1972). Circannual The stages of the ovarian cycle of the rhythms in some species can be affected by American Goldfinch at Guelph bore a similar day-length, which strongly suggests that it is relationship to the testicular cycle as that in the annual fluctuations of the photoperiod the European Goldfinch (Car&&is carduelis) that maintain the timing of the rhythm (Gwin- in Australia (Middleton 1971). The oviduct ner 1973, 1975). For the goldfinch the precise cycle closely follows the ovarian cycle, with relationship that exists between events of the peak oviduct weight coinciding with the peak annual cycle implies that some internal rhythm of egg-laying in the nesting population (Mid- is present; while the precision with which dleton in press). events occur each year at Guelph suggests a The most striking feature of the annual response to a very predictable environmental cycle of the goldfinch at Guelph is the amount timer, the most likely being photoperiod (Lofts of time devoted to molt (Fig. 3). Although and Murton 1968, Morrison and Wilson 1972, this is mainly due to the gradual nature of Immelman 1973, Follett and Davies 1975, the molt in the head region, associated with Lofts 1975). In these respects the goldfinch the emergence of a sexually dimorphic plum- appears typical of temperate zone passerines, age in both first-year birds and old birds but this does not explain what regulates its (Middleton 1977), the peak period of both late gonadal development. prenuptial and postnuptial molts lasts longer Mundinger (1972) showed that the dura- than is characteristic of temperate zone pas- tion of the regeneration, acceleration and serines (Middleton 1977). The absence of culmination phases of the testicular cycle of molting during the height of the nesting the goldfinch resembled that in four other season agrees with most temperate zone spe- species, but the goldfinch differed in the late cies in which breeding and molt are usually timing of all its phases. Mundinger noted the mutually exclusive events (Davis 1971, Payne prolonged prenuptial molt but made no further 1972, Morton and Welton 1973). By contrast, reference to it. In two incomplete experiments molt apparently overlaps with migration in with goldfinches (most birds died from coc- both spring and fall. Usually molt is either cidiosis early in the experiments), I adjusted completed before, or is suspended during, mi- the photoperiod in an attempt to gain some in- gration in most temperate zone species (Hau- sight into the proximate factor that controls kioja 1971, Payne 1972, Pearson 1973) be- the gonadal cycle (Middleton, unpubl. data). cause the energetic demands of the two All birds that survived the experiments molted processes are incompatible ( Evans 1969). before showing a gonadal response (assessed ANNUAL CYCLE OF THE AMERICAN GOLDFINCH 405

by laparotomy and bill pigmentation; see The cycles of the seminal sacs and oviducts Middleton 1971, Mundinger 1972). These closely followed those of the gonads. tentative results suggest that molt was initi- Both prenuptial and postnuptial molts were ated by a changing photoperiod and that prolonged and occupied most of the year. once it had reached a certain stage, gonadal The nesting season was the only time when recrudescence began. Thus, molt, rather than molt was not observed. the gonadal cycle, may be the process that is Based on banding data, migration occurred initially triggered by a photoperiodic response in early May to mid-June and between late in the goldfinch. This may also be true for October and mid-December. some other temperate zone passerines (Snow The timing of all events in the annual cycle 1969, Dolnik 1975). is assessed. Precision in the timing of the Although the exact nature of the interac- gonadal development, molt and migration is tion between molt and reproduction is poorly noted, which suggests a response to a pre- understood, the hormones associated with each dictable proximate factor, the most likely be- process usually inhibit each other in temperate ing photoperiod. As molt occupies the greater zone species (Lofts and Murton 1968, Payne part of the year, the timing of events within 1972). The hormonal balance is probably the annual cycle of the goldfinch appears to different during the prenuptial and postnup- be controlled by molt and not by the gonadal tial molts in species that molt twice a year cycle. ( Payne 1972). In the goldfinch, rapid gonadal growth does not occur until May, when the ACKNOWLEDGMENTS prenuptial molt is well advanced (Middleton 1977). Thus the incorporation of the prenup- Thanks are offered to Murray Pengelley, Glen Fox, Marv and Mike Dyer. Richard Hamilton. and Alex tial molt into the annual cycle may prevent Derry, for their assistance in the field. E: D. Bailey early gonadal recrudescence, and may be the and D. E. Joyner provided valuable criticisms of factor which ensures the late timing of the con- earlier drafts of the paper, while the comments of stituent phases of the gonadal cycle discussed P. Mundinger and an anonymous referee further by Mundinger ( 1972). strengthened it. The research was financially sup- ported by the National Research Council of Canada, Clearly, molt is the most time consuming through grant #A3911 to the author. event in the annual cycle of the goldfinch at Guelph. It may therefore be the basic event LITERATURE CITED which anchors the timing of the other major events. If so, the annual cycle of the American DAVIS, J. 1971. Breeding and molt schedules of the Rufous-collared Sparrow in coastal Peru. Goldfinch is apparently similar in its physio- Condor 73: 127-146. logical timing to that proposed for some trop- DOLNIK, V. R. 1975. Photoperiodic control of sea- ical species (Miller 1962, Snow and Snow sonal cycles of the body weight, moult and sexual 1964, Lofts and Murton 1968, Snow 1976), activitv in Chaffinches ( F&u&z coelebs). Zool. Zh. 5&1048-1056. ’ - and may be unusual for a temperate zone spe- ERSKINE, A. J. 1971. Growth, and annual cycles in cies. Although Mayr and Short (1970) sug- weights, plumages and reproductive organs of gested that this species has no obvious close Goosanders in eastern Canada. Ibis 113:42-58. relatives, my results suggest that it may have EVANS, P. R. 1969. Ecological aspects of migra- tropical affinities as suggested for the other tion, and pre-migratory fat deposition in the Lesser Redpoll Carduelis flammea cabaret. Con- North American “goldfinches,” C. psaltria and dor 71:316-336. C. lawrencei (Mayr and Short 1970). FOLLETT, B. K. AND D. T. DAVIES. 1975. Photo- periodicity and the neuroendocrine control of re- SUMMARY production in birds. Pp. 199-224. In M. Peaker [ed.], Avian physiology. Academic Press, Lon- The gonadal and molt cycles of the American don. Goldfinch were studied at Guelph, Ontario FORBUSH, E. H. Ah-D J. B. MAY. 1939. Natural history of the birds of eastern and central North between 1968 and 1975. The gonads of 200 America. Houghton Mifflin Co., Boston. birds were examined while molt data were ob- GWINNER. E. 1973. Circannual rhvthms in birds: tained from an additional 3,804 individuals Their interaction with circadian chythms and en- trapped throughout the year. vironmental photoperiod. J. Reprod. Fertil. 19: Reproductive condition was first reached Suppl., 51-65. GWINNER, E. 1975. Circadian and circannual rhythms by mid-June, and all birds were in breeding in birds. Pp. 221-285. In D. S. Farner and J. R. condition by mid- July. Breeding condition King [eds.], Avian biology. Vol. 5. Academic persisted until late August. Old male gold- Press, New York. HAUKIOJA, E. 1971. Summer schedule of some finches had heavier testes, and reached breed- subarctic passerine birds with reference to post- ing condition earlier, than first-year birds. nuptial moult. Kevo. Subarctic Res. 7:60-69. 406 ALEX L. A. MIDDLETON

HOLCOMB, L. C. 1969. Breeding biology of the growth in Tree Sparrows (Spizella arborea). American Goldfinch in Ohio. Bird-Banding 40: Auk 89: 146-155. 26-44. MORTON, M. L., AND D. E. WELTON. 1973. Post- I~~ELMAN, K. 1971. Ecological aspects of periodic nuptial molt and its relation to reproductive reuroduction. PD. 341-389. In D. S. Famer and cycle and body weight in Mountain White- J.R. King [eds.], Avian biology. Vol. 1. Academic crowned Sparrows (Zonotrichia leucophrys ori- Press, New York. antha). Condor 75: 184-189. IMLII\IELMAN, K. 1973. Role of the environment in NEWTON, I. 1972. Finches. Collins, London. reproduction as source of “predictive” informa- NICE, M. M. 1939. Territorial song and non-terri- tion. Pp. 121-147. In D. S. Farner red.], torial behavior of Goldfinches in Ohio. Wilson Breeding biology of birds. Natl. Acad. Sci., Bull. 51:123. Washington, D.C. PAYNE, R. B. 1969. Breeding seasons and repro- LOFTS, B. 1975. Environmental control of reproductive physiology of Tricolored Blackbirds and duction. PO. 177-197. In M. Peaker Jed.1. Avian Redwinged Blackbirds. Univ. Calif. Publ. Zool. physiology: Academic Press, London. - ’ 90: 1-115. LOFTS, B. AND R. K. MURTON. 1968. Photoperiodic PAYNE, R. B. 1972. Mechanisms and control of and physiological adaptations regulating avian molt. Pp. 103-155. In D. S. Farner and J. R. breeding cycles and their ecological significance. King [eds.], Avian biology. Vol. 2. Academic J. Zool. Lond. 155:327394. Press, New York. LOFTS, B. AND R. K. MURTON. 1973. Reproduction PEARSON, D. J. 1973. Moult of some palaearctic in birds. Pp. l-107. In D. S Farner and J R. warblers wintering in Uganda. Bird Study 20: King reds.], Avian biology. Vol. 3. Academic 24-36. Press, New York. SCOTT, D. M. AND A. L. A. MIDDLETON. 1968. The MARSHALL, A. J. 1959. Internal and environmental annual testicular cycle of the Brown-headed control of breeding. Ibis 101:456-478. Cowbird (lMolothrms ater). Can. J. Zool. 46:77- MARSHALL, A. J. AND C. J. F. COOMSS. 1957. The 87. interaction of environmental, internal and be- SNOW, D. W. 1969. The moult of British thrushes havioral factors in the Rook, Corvus f. frmgilegus and chats. Bird Studv 16:115-129. Linnaeus. Proc. Zool. Sot. London 128:545- SNOW, D. W. 1976. The relationship between cli- 589. mate and annual cycles in the Cotingidae. Ibis MAYR, E. AND L. L. SHORT. 1970. Species taxa of 118:36&401. North American birds. Nuttall OrnithoI. Club. SNOW, D. W. AND B. K. SNOW. 1964. Breeding Publ. No. 9. seasons and annual cycles of Trinidad land birds. MILLER, A. H. 1962. Bimodal occurrence of breed- Zoologica 49: l-39. ing in an equatorial sparrow. Proc. Natl. Acad. STOKES, A. w. 1950. Breeding behavior of the Sci. 48:396-400. Goldfinch. Wilson Bull. 62: 107-127. MIDDLETON, A. L. A. 1971. The gonadal cycle of SUTTON. G. M. 1959. The nesting frinaillids of the the Goldfinch in south-eastern Australia. Emu Edwin’ S. George reserve, soutbeaste& Michigan 71: 159-166. (Part 2 ). Jack-pine Warbler 37 :3750. MIDDLETON, A. L. A. 1972. The structure and pos- TRAUTMAN, M. B. 1940. The birds of Buckeye sible function of the avian seminal sac. Condor Lake, Ohio, Univ. Mich. Mus. Zool. h4isc. Publ. 74: 185-190. 44. MIDDLETON, A. L. A. 1974a. Age determination in TYLER, W. M. 1968. Spinus t&is tristis (Lin- the American Goldfinch. Bird-Banding 45:293- naeus) Eastern American Goldfinch. Pp. 447- 296. 466. In 0. L. Austin Jr. ted.], Life histories of MIDDLETON, A. L. A. 1974b. Spermiation and sperm North American cardinals, grosbeaks, buntings, transport in passerine birds. J. Reprod. Fertil. towhees, finches, sparrows and allies. U.S. Natl. 40:31-37. Mus. Nat. Hist. Bull. 237, Part 1. MIDDLETON, A. L. A. 1977. The molt of the Ameri- VOX HAARTLIAN, L. 1972. Further observations on can Goldfinch. Condor 79 :440444. late nesting in cardueline finches. Omis Fenn. MIDDLETON, A. L. A. Influence of age and habitat 49:15. on reproduction by the American Goldfinch. WISEMAN, A. J. 1975. Changes in body weight of Ecology. In press. American goldfinches. Wilson Bull. 87:390-411. MUNDINGE~, P. C. 1972. Annual testicular cycle and bill color change in the Eastern American Department of Zoology, University of Guelph, Guelph, Goldfinch. Auk 89:403-419. Ontario, Canada NlG 2Wl. Accepted for publication MORRISOI\T,J. V. AND F. E. WILSON. 1972. Ovarian 13 February 1978.