Alloparental Care in the Purple Gallinule

Condor, 80:382-390 0 The Cooper Ornithological Society 1978

ALLOPARENTAL CARE IN THE PURPLE GALLINULE

CHARLES ONEIL’ KREKORIAN

Purple Gallinules ( Gallinula mwtinica) oc- The objectives for this study of Purple Gal- cupy lowland fresh-water marshy habitats linules in Costa Rica were: (1) to establish throughout the Americas (Wetmore 1965). whether they have alloparental care, and to Their breeding range extends from central learn the status or condition of the helpers; Ohio in the United States to northern Argen- (2) to describe the types of assistance pro- tina and Uruguay in South America (A.O.U. vided, and quantify the relative contributions 1957). With the exception of Gross and van made by various members of the social unit, Tynes’ ( 1929) excellent description of the nest, including the chicks; and (3) to gain infor- eggs, and nesting behavior of the Purple Gal- mation about Purple Gallinule social struc- linule on Barro Colorado Island, Panama, and ture and data that might serve for evaluating a description of its nesting behavior in Ohio the effect of helpers on the reproductive suc- (Trautman and Glines 1964)) little is known cess of the breeders. about the ethology of this species. Available information characterizes the Purple Gallinule as monogamous, with the parents setting up STUDY AREA feeding territories and sharing care of the pre- This study was conducted on the shallow 7-acre pond cocial chicks (Gross and van Tyne 1929, Mean- (2.83 ha) located on the grounds of the Centro ley 1963, Trautman and Glines 1964). Allopa- Agronomic0 Tropical de Investigation y Ensenaiiza rental care (“helpers”) in this species is not (CATIE ) near Turrialba, Costa Rica. The pond, its flora and fauna, and the climate of the area were de- mentioned in the literature although some scribed in detail by Jenni and Collier ( 1972). zoologists (Gerald Collier and Paul Fromer, The distribution of vegetation on the pond varied pers. comm.) have suspected it. “Alloparental” from year to year due to CATIEs’ management prac- is a neutral term used to describe the behavior tices. During my study, three distinct zones of vegetation existed, each characterized by one or more of members in a social group other than the conspicuous plant species (Fig. 1). The first con- parents which assist in the care of offspring sisted of papyrus (Cyperus papyrus) which attained (Wilson 1975). a height of 3.5 m and grew on the north and east Approximately 80 species of birds in 32 dif- sides of the pond. The second zone was made up of ferent families are known to exhibit coopera- semiaquatic grasses and sedges, 0.3-2 m high, which covered the west and the east ends of the pond. The tive breeding in which helpers feed nestlings grasses (Leersia hexandru) and sedges (Eleochuris geni- and occasionally fledglings (Skutch 1961, Har- culutu) also covered mud islands ( 5-20 m diam. ) at rison 1969, Fry 1972, Grimes 1976, Rowley the west and north parts of the pond and a large 1976, Woolfenden 1976, Zahavi 1976). In ad- island (ca. 35 m x 40 m) on the southeastern half of the pond (Fig. 1). The grasses occupied about dition, some helpers aid in nest and/or terri- 35% of the pond area. The third zone, dominated by tory defense (Ligon 1970, Parry 1970, Brown broad-leaved water lilies (Nymphueu sp. ) though 1972, Zahavi 1974, Woolfenden 1975), and often mixed with other low lying aquatic plants, nest sanitation (Skutch 1961, Ligon 1970, Fry covered most of the central and remaining portions 1972). Nest building, incubation, and brood- of the pond where grasses were not found. The pro- file of the third zone was low and open, with grassless ing by helpers have been recorded less fre- mud islands scattered throughout. A small area near quently but do occur in “mutual helpers” such the center of the pond was open water. as the anis (Crotophagu; Vehrencamp 1977) and in some other groups such as bee-eaters MATERIALS AND METHODS (Meropidae; Fry 1972) and swallows (Hir- undinidae; Skutch 1961). Observations were conducted for 117 days usually between 06:30-lo:30 from 1 October 1976 to 24 Alloparental care is currently of sociobiolog- February 1977 using a 15-60~ zoom telescope and ical interest because it is used as evidence of 10 x 50 mm binoculars. The six families studied kin selection and altruism (Wilson 1975). The were observed a total of 238 h. Five hours of observa- controversy over kin selection versus individ- tions were made between 13:00-17:OO when weather conditions were cool, and the birds active. ual selection (Hamilton 1964) can be resolved Twenty-nine Purple Gallinules were trapped with only when additional studies of alloparental self-tripping drop cages (50 x 50 x 24 cm) made of care document the filial relationship of helpers coarse mesh hardware cloth on a wood platform. Platforms were placed in the pond near the shore and to breeders and the effect of helpers on re- next to the grass islands frequented by each family. production. Captured gallinules were weighed, measured (shield, [3821 ALLOPARENTAL CARE IN GALLINULES 383

~rasr island

mud island

I I I I bamboo island 0 10 20 30 m

FIGURE 1. The pond at Turrialba, Costa Rica, showing the distribution of vegetation during the study and the approximate location of Purple Gallinule family territories. beak, and leg) and banded with a USFWS metal rather than plumage condition. Since my use of the band and a unique combination of at least two colored term “chick” is based on behavior rather than plum- plastic bands using red ( R ), blue ( B ), yellow ( Y) , age, it does not always denote young in natal down. and orange (0) which were placed on the lower I use the term “chick” here to denote any individual tibiotarsi. With the bird facing the observer the receiving food from an adult or older juvenile. Thus, bands were recorded in sequence from top to bottom, this category included individuals up to 60 days from the observers’ right to his left. In recording the old because they received help from adults even color code a dash always follows any bands on the though they wore the juvenile plumage described right leg and precedes any bands on the left leg. In by Wetmore ( 1965). all cases, observations of families with chicks began before all members, including chicks, were banded. JUVENILE PLUMAGE Three stages in the life history of the Purple Gal- linule (chick, juvenile, and adult) generally are Head, neck, sides, and tibia buffy brown; back, recognized. Each stage is associated with a charac- rump, and tail dull brown; wings greenish-blue washed teristic appearance, but the history of molts and plum- with brown; throat, breast, abdomen, and under ages is not known in detail. Nevertheless, the follow- tail coverts white. I considered individuals to be ing descriptions, modified from Wetmore ( 1965), juveniles (9-24 weeks of age) until they had full allowed stage identification of the socially interac- adult plumage and markings. tive birds. ADULT PLUMAGE DOWNY CHICKS (Sexes alike.) Head, breast, sides, and wings deep Above black, underneath brownish black; crown, blue; back, rump, tertials, and tail dull green; abdo- sides of head from bill to back of eyes, and throat men and tibia black; under tail coverts white. with filaments of very pale bluish white. Fully mature adult Purple Gallinules have a con- Traces of brown began to appear on the black spicuous light blue frontal plate, and the bill is plumage of some chicks when they were approxi- bright red at the base and yellow over the distal mately 2 weeks old. When they were 34 weeks old, 12-20 mm. the ventral half of the body was light beige-brown The time and duration of feeding and other and the dorsum still black with a greenish-olive pertinent behaviors were recorded with a stopwatch. tint. Juvenile plumage was attained sometime be- Each time a chick fed or was being fed by a juvenile tween 5 and 7 weeks of age by some chicks. In this or adult, the identity of individuals involved (based paper, chick status is based on a behavioral criterion on the band combinations) was recorded. When the 384 CHARLES ONEIL’ KREKORIAN

TABLE 1. Purple Gallinule families studied and the number of times (numbers in parentheses) banded family members were observed to feed chicks. Feedings of chicks by unbanded adults [UA] are also included. Additional family growth after March 1977 is included in the text.

Family 1 Adult O-O (46), B-B (23), Y-Y (39) Juvenile R-R (37), B-O (7) Chick B-R, B-Y, R-O Family 2A Adult Y-R (8), BB-0 (13), BB-Y (4), UA(s) (2) Juvenile R-Y (7), Y-O (26) Chick Y-B, BB-R Family 2B Adult Y-R (9), BB-0 (13), BB-Y (18), R-Y (27), 00-Y (O), Y-O (0), UA(s) (14) Juvenile Y-B (4), BB-R (6) Chick RR-O, 00-R Family 3 Adult O-Y (7), RR-Y (5) Juvenile None Chick RR-B, YY-B, YY-R Family 4 Adult R-B, O-B, YY-0, 0-BB Juvenile None Chick 3 unbanded Family 5 Adult 4 unbanded Juvenile 2 unbanded Chick 3 unbanded Family 6 Adult 2 unbanded Juvenile 2 unbanded Chick 3 eggs bands could not be seen or the individuals involved Family 1 captured on 14 and 15 October had no bands, the birds were designated as chick, were: BB (306 g), O-O (263 g), Y-Y (228 juvenile, or adult and the performed behavior described. Territory defense, conspicuous postures, and g), and juvenile R-R (215 g). Juvenile B-O food items were recorded and described. weighed 230 g on 25 October. The base of the bill of both B-B and O-O was bright red, RESULTS which is typical of fully mature adults, but the base of Y-Ys’ bill was light red. The base FEEDING OF CHICKS BY FAMILY MEMBERS of Y-Ys’ bill had become bright red by 2 Feb- Three families of Purple Gallinules were ruary 1977. The lighter weight and the bill studied in detail and less detailed observa- coloring of Y-Y compared to other adults at tions were made on an additional three fam- capture indicate that it had recently acquired ilies. Families studied in detail ranged in adult plumage. The similarity in weight of Y- size from five individuals in Family 3 to ten Y and the juvenile B-O, which was caught banded individuals in Family 2B (Table 1). 10 days after Y-Y, also supports this conclu- Family size and the number of individuals sion. banded are the same except for Family 2 in Data for Family 2 are presented in two which at least one individual was not banded. parts because there were two pairs of chicks. Four families had easily distinguishable ju- Y-B and BB-R hatched on 23 November 1976 veniles; the other two lacked juveniles (Table ( *l day), and these chicks along with the 1). Alloparental feeding occurred in all fam- adults and juveniles who fed them are desig- ilies with both juveniles and chicks (3 of 6 nated Family 2A. The second pair of chicks families). (RR-O and 00-R) was sighted 28 January Family 1, located in the southeastern part of 1977 when they were estimated to be 2 or 3 the pond (Fig. l), consisted of eight individ- weeks old. These chicks with the adults and uals, including three chicks. All members of juveniles who fed them are designated as the family fed the chicks (Table 1). The two Family 2B. Both pairs of chicks are presumed juveniles in Family 1 fed the chicks 70 times to have been in the same family for two rea- or 13% of the observed feeding instances sons. First, all of the chicks foraged over and (Table 2). The three adults accounted for frequented the same areas within Family 2s’ 26% of the feeding instances observed. On the territory (Fig. 1). Second, the two chicks of average, each juvenile of Family 1 fed the the second brood were fed by the two chicks chicks about as frequently as did each adult of the first brood and also by four of the six (Table 2). individuals which fed the first pair of chicks The weights of the adults and juveniles in (Table 1). ALLOPARENTAL CARE IN GALLINULES 385

I saw two juveniles, three marked adults, TABLE 2. The percentage of feeding instances that and an unmarked adult feed the chicks of self-feeding and allo-feeding in chicks was observed Family 2A during the 9 weeks they were in two families of Purple Gallinules. Families 2A and 2B refer to the first and second pairs of chicks pro- watched. Because I also saw at least two, duced by Family 2. The numbers of chicks, juveniles, possibly three, unmarked adults in this area and adults are shown in parentheses. concurrently, the chicks may have been fed by two different unmarked adults (Table 1). Familv 1 2A 2B The two juveniles of Family 2A (R-Y, Y- Total feeding instances 550 469 210 0) accounted for 9% of the total feeding in- Total chick ;; i;{ 7; {;I 55” i;j ’ stances of the two chicks, while the four (or Total juvenile five) adults accounted for 18%. On the aver- Per juvenile Total adult age, the juveniles of Family 2A fed the chicks 2& I& 4??5j Per adult 8.7 4.5 8.2 frequently as did the adults (Table 2). When the chicks of Family 2B were seen on 28 January, the chicks of Family 2A were 67 at 1 days old and in juvenile plumage. seen only on the eastern half of the island and The next day, these two juveniles fed the new in other areas to the east of the island (Fig. chicks whose age at that time I estimated to 1). I never saw fighting between Families 1 be between two and three weeks. All to- and 3. gether, the chicks of Family 2B were fed by Both adults fed the chicks, providing food the two juveniles, and at least five adults. to the three chicks with about the same fre- One of these adults (R-Y) had recently ac- quency (Table 1). The body weights of quired adult plumage and had been a juve- the adults were 275 g (O-Y) and 256 g nile when observations were made on Family (RR-Y). On the basis of its heavier body 2A. An unbanded adult fed the chicks 14 weight, I assume that O-Y was the male and times (Table 1) during the three weeks I RR-Y was a female. Data from Collier (pers. observed Family 2B. Again, the chicks may comm.) and Wetmore (1965) support this have been fed by more than one unbanded assumption. adult because several unbanded adults were Three new downy chicks were sighted by present in this area. One adult which was Reeves at different times, beginning 29 May caught and banded (00-Y) on 20 February 1977, with adults O-Y, RR-Y, and the juve- 1977, fed the juvenile Y-B (chick of Family nile YY-R who had recently acquired its ju- 2A), who in turn fed this food to two new venile plumage. These chicks were fed by downy chicks on 11 April 1977 (L. Reeves, YY-R (Reeves, pers. comm.) which I banded pers. comm.). This third pair of chicks, mem- as a chick 22 February 1977. bers of Family 2C, frequented the same areas Families 4-6 lived on the north side of the as other chicks in Family 2, and was fed by pond. Distance and vegetation made ob- 00-Y, the adult R-Y, the juvenile Y-B, and servations from the south side of the pond and an unbanded adult (Reeves, pers. comm.). banding difficult (Fig. 1). It was possible, Thus, the number of gallinules in this family however, to observe care of the chicks by is not known but could be 13 or 14. adults and juveniles (when present) in Fam- In Family 2B the new juveniles (Y-B, BB- ilies 4 and 5 when the chicks were in the third R) accounted for only 5% of the total feeding zone of vegetation or on the shore of grass- instances for the chicks (RR-O, 00-R)) while covered islands. During these periods, I adults accounted for 41%. The percentage of counted the number of chicks, juveniles, and feedings contributed per individual by adults adults present in the area and recorded the (8.2%) was three times that by each juvenile number of different individuals observed (2.5%) (Table 2). feeding the chicks. I assumed that Purple Family 3 consisted of two adults and three Gallinules who fed particular chicks were chicks when observations began. Its area of in the same family because family members activity included the western half of the large excluded individuals of other families from grass island. Previously, this area had been their territories. As family 6 had no chicks occupied exclusively and defended by Fam- when I terminated my observations, but rather ily 1; but on 19 January 1977, O-Y and two a nest containing three eggs, I assumed that unbanded adult gallinules, one of which adult and juvenile gallinules not chased from carried nesting material in its bill, were ob- areas near the nest were members of this served on the western shore of the grass is- family. Pertinent observations of these fam- land. From that time onward, Family 1 was ilies are recorded in Table 1. 386 CHARLES ONEIL’ KREKORIAN

FEEDING DURATION Chicks began pecking at the lily fruit during The duration of feeding bouts in which a ju- this phase but usually were unsuccessful in venile helper fed a chick varied from approxi- removing any of it. They never were seen to mately 10 s to 10 min. Brief bouts usually in- procure a lily fruit from below the water or volved a juvenile bringing a single food item mud surface or to successfully prepare one for to a chick and then leaving. Long bouts oc- eating. Before a lily fruit was eaten, adults curred when one or more chicks were fed removed at least a portion of the outer, black, pieces of lily fruit by the helper, while all shell-like covering. This often took l-2 min birds were gathered around the lily fruit. and sometimes longer. Chicks also took sting- During these bouts, the chicks were sometimes less bees, unidentified insects, and spiders fed more than 10 times from the beak of the from the air and lily flowers, and the seeds of helper. Juvenile helpers commonly fed chicks grasses. for 2-5 min when they were gathered at a The third phase of feeding interaction be- lily fruit. tween chicks and other family members began at seven weeks and continued until chicks CHANGES IN THE MODE OF were approximately eight weeks old. Dur- FEEDING WITH TIME ing this period, adults and juveniles secured lily fruit, sometimes preparing it, and gave Feeding interactions between helpers and it to chicks who quickly swallowed it or took chicks changed considerably as the chicks it to another place within the territory to eat. grew older. I recognize three phases based Adults and juveniles exhibited no aggression on my observations of Families 1, 2, and 3. toward the chicks and made no attempt to The intervals covered by each phase are ap- retrieve the lily fruit that the chicks carried proximate as I did not observe the exact date off. Twice I saw the same adult provide a lily of hatching of most chicks. fruit for two chicks in succession before secur- Phase 1 lasted from hatching until the ing one for itself. Beak-to-beak feeding be- chicks were 3-5 weeks old. During this period, tween a chick and an adult or juvenile oc- the chicks typically followed one or more curred less frequently during this phase than adults and/or juveniles to a location and earlier. At this time I also saw a 45day-old waited while family members gathered food chick in Family 2A (BB-R) prepare a lily fruit nearby or at distances 3040 m away. Once for the first time by removing the outer black a family member had food, it either walked covering, and during the following week it or flew to the chick or chicks and fed them prepared more lily fruits. I first saw a chick by passing the food from its beak to the in Family 1 prepare a lily fruit at approxi- chicks’ beak. The chicks pecked at the beak mately 51 days of age. I consider the third of the juvenile or adult at this time and fre- phase to have ended when juveniles and adults quently showed a “begging posture” when the stopped feeding the chicks from the beak. The family member approached. In this posture chicks in Family 2A were last seen fed from the neck of the chick is stretched forward and the beak of an adult when they were 57 days lowered below the level of the rump. At the old. The last time a juvenile in Family 2A same time the wings are elevated and often was observed to feed a chick from the beak vigorously moved up and down synchronously. was when the chick was 54 days old. On the During this performance, I frequently heard basis of age estimates of the chicks in Family the chick peeping. Chicks received both plant 1, adults stopped feeding chicks from the beak and animal food, including stingless bees when they were about 59 days old, while ( Trigonu sylvuestriana and T. corvimz), other juveniles stopped when chicks were about 49 insects, and pieces of water lily fruit during days old. However, I saw chicks in both Fam- phase 1. ily 1 and 2A take a lily fruit from an adult During phase 2, which began when the when they were about 60 days old. The chicks chicks were approximately five weeks old and became independent at about nine weeks of continued until they were about seven weeks age. old, they more actively solicited food from their elders. Chicks walked or ran to family FOOD EATEN members feeding on lily fruit out in the pond and were fed from the beak of the helper or Lily fruit, which was very abundant because parent as it chipped off pieces of the fruit. of CATIEs’ pond, was the food most fre- They also ate fallen fragments of fruit and oc- quently eaten by Purple Gallinules. Seventy- casionally quickly swallowed an entire small one percent of the food items positively iden- lily fruit prepared by an adult or juvenile. tified at the time of eating were lily fruits. ALLOPARENTAL CARE IN GALLINULES 387

In addition, stingless bees ( 16% ), frogs (7% ), Another time, the juveniles of Family 1 be- grass seeds (2%; L. h#exandra on islands and gan fighting two unmarked adult Purple Gal- Paspalum virgatum on the shore), spiders, linules near the western border of Family ls’ worms, and fish were eaten. Nineteen frogs territory (prior to establishment of Family were taken by Families 1, 2A, and 2B. Each 3). Two adults from Family 1 flew to the site time a frog was caught, most or all members of the fight, the fight soon stopped, and the of the family tried to obtain a piece. Two or two intruders left the area. more individuals often pulled at opposite ends of the frogs’ body. Individuals who ran FAWLY RELATIONSHIP OF HELPERS off with a piece of the frog were chased by Most of the individuals in four of the families family members who had been unable to se- studied were banded, and three of the fam- cure a portion for themselves. If one bird ilies were sufficiently small and geographically was successful in obtaining anothers’ piece, contained to provide evidence regarding the the roles of chaser and chased were reversed. parentage of the chicks and/or the filial re- On ten occasions this activity persisted for at lationships among helpers, parents, and chicks. least 8-23 min, when I ceased watching. Copulation between adults would be a better The chicks accompanied members of the indicator of mateship and chick parentage family most of the time as they moved through than close association. However, during my their family territory. Thus, inadvertently or study I witnessed copulations only twice. Even otherwise, helpers assisted the young in recog- when observed, copulation does not prove nizing and perhaps preparing food items, whose sperm fertilizes the eggs, and thus especially lily fruit. parentage and filial relationships between family members cannot always be known with ADDITIONAL FORMS OF HELP certainty. The juveniles and chicks in Family Purple Gallinule helpers (both adults and ju- 1 were presumed to be the offspring of the veniles) also participate in territory and chick adults O-O and B-B because the third adult defense (see Fig. 1 for the approximate loca- in Family 1 (Y-Y) had recently acquired tion and size of family territories). Juveniles adult plumage. Thus, the juveniles were prob- of Families 1 and 2 performed some form of ably full siblings of the chicks. The two juve- defense a total of 18 times during the 213 h niles (R-R and B-O) were never seen to feed they were observed. Seventy-two percent of individuals other than the chicks. The adults these defensive actions involved territorial de- of Family 3, O-Y and RR-Y, were seen with fense and 28% chick defense. Forty-four per- three new chicks in May 1977. These chicks cent (8 of 18) of the defensive actions were were fed by YY-R, who had recently acquired directed against conspecific birds and 56% juvenile plumage. Since YY-R had been one (10 of 18) against birds of other species. Most of three chicks in the first brood cared for ex- of the heterospecific defense involved Jacanas clusively by adults O-Y and RR-Y (no ju- (Jacana spinosa) while the rest was against veniles were present in this family when Common Gallinules ( Gal&&a chloropus) . study began) this is the strongest circumstan- As all three adults in Family 1, and three tial evidence possible for demonstrating its adults in Family 2B were observed in some sibling relationship to the new chicks. The form of defense, I assume that other non-par- helpers in Family 4 also may have been sibs ent adult helpers also participated in defense. of the chicks. Two of the adult-plumaged gal- The adults, however, have been excluded linules in this family weighed less than 206 from these totals because of the difficulty in g and had just acquired their adult plumage. knowing the exact parentage of chicks. Nest The parents of the juvenile helpers in Fam- defense also may occur, but data are lacking ily 2A and 2B are unknown. Three banded as is information on the role of helpers in nest adults (Y-R, BB-0, and BB-Y) and at least building, incubation, and brooding. one unbanded adult fed the chicks of Family On three occasions noisy fights between 2A (Table 1). These same banded adults, two or more individuals from each of two another adult with recently acquired plumage families occurred. Other members of both ( R-Y), and at least one unbanded adult also families gathered at the site of the fight so fed the chicks of Family 2B (Table 1). Be- that as many as ten adults could be seen at cause at least four mature adults of unknown one time. Twice these conflicts involved sex fed the chicks of Families 2A and 2B, Families 2 and 5 where their territories were any combinations involving two of the adults adjacent. The fighting individuals repeatedly were potential parents of the first and second flapped their wings and kicked each other. pair of chicks. Thus, it is not possible to know 388 CHARLES ONEIL’ KREKORIAN which two adults (assuming monogamy in Purple Swamphens in New Zealand also this species) were the parents of the chicks breed in pairs and groups of 3-6 individuals and/or the juveniles that fed the chicks. Be- (Craig 1975). Copulation is a group event cause of the small age difference (about 49 with most copulations involving more than days) between the first (Family 2A) and two adults. In groups, one or more females second (Family 2B) pairs of chicks produced lay in the same nest at the same time, with the by Family 2, some of the helpers may not number of eggs being variable. All birds in have been the offspring of the parents of the group, including the older chicks, care Family 2B. for the young chicks. All eight Purple Gallinule broods produced DISCUSSION at CATIE during this study contained two or three chicks and a nest contained three eggs. Alloparental care in this population of Purple Purple Gallinule nests in Panama also con- Gallinules was common. Of the six families tain three or four eggs (Gross and van Tyne studied, five had helpers (Families 1, 2, 3, 4, 1929). This consistency in clutch and brood and 5), and one (Family 6) had potential size would seem to rule out the type of mating helpers (two unbanded juveniles) for a clutch system found in the Purple Swamphen. The of 3 eggs. Both adults and juveniles helped large size of Family 2 (Table 1) and the supposed parents feed and defend the young, closeness in age of the three separate pairs of and defend the family territory. In Families offspring produced suggest several other pos- 1 and 2A each juvenile fed the chicks nearly sibilities in the type of mating system in ad- as frequently on the average as each adult. dition to that found in the Tasmanian Native Scant knowledge exists on the breeding Hen. Some families may contain more than biology of almost all rails, but the Purple Gal- one breeding pair. Polygyny and promiscuity linule is the eighth of 132 species of Rallidae are also possible but seem less likely. The ab- for which helpers have been documented. sence of conspicuous sexual dimorphism in Alloparental behavior has also been observed this species makes it difficult to determine the in: the Tasmanian Native Hen (Gallin& sex ratio of large groups. Even though males mortierii) and the Dusky Moorhen (G. tene- are heavier than females, weights cannot be brosa; Ridpath 1972a, b), the Purple Swamp- judged from a distance, and unless the ages hen (Porphyrio porphyria; Craig 1975), the of all adults in a group are known, it is not Common Gallinule or Moorhen (Grey 1927, possible to tell the sex of most individuals be- McIlhenny 1934, Skutch 1961,), the European cause some birds may still be growing. Such Coot ( FuZi,caatru; Ruthke 1939)) the Red-and- information is required to determine the type White Crake ( Laterallus Zeucopyrrhus; Meise of mating system and the exact filial relation- 1934), and the Black Crake (Porxanu fluvi- ship of the helpers to the breeders and chicks. rostra; Brooke 1975). Helpers of some chicks at CATIE, for example Although most of the above findings were those of Family 2, could be their “grandpar- based on limited observations, they suggest ents,” “aunts” and/or “uncles.” that helpers may occur regularly, or at least Studies of alloparental care that have con- not uncommonly, in these species. They also tinued long enough to determine the filial re- suggest that alloparental care may be wide- lationship of the helpers to the breeders have spread in this family. The kind and amount usually found that the helpers are offspring of help individuals may provide remains to of the breeders and siblings of the chicks be documented for most of the species dis- (Rowley 1965, Fry 1972, Ridpath 1972a, cussed above. At present, quantitative infor- Parry 1973, Craig 1975, Woolfenden 1975). mation on helpers in the Rallidae is available Rarely are helpers more distantly related. only for the Tasmanian Native Hen, the Pur- However, Woolfenden (1975) found that four ple Swamphen, and the Purple Gallinule. male Florida Scrub Jays (Aphelocoma coeru- Three of the families in my study had only Zescens) (uncles) helped care for the off- two adults when observations began. An ad- spring of their brothers and their mates. Cir- ditional family (Family 1) had three adults, cumstantial evidence from Families 1, 2, and one of whom (Y-Y) had recently acquired 3 strongly suggests that some Purple Gallinule adult plumage. These observations suggest helpers are older siblings from earlier broods, that the Purple Gallinules at CATIE are but helpers may also have more distant filial monogamous, like those studied elsewhere. relationships. But, the possibility of simultaneous poly- Evidence that helpers increase reproductive andry, as exhibited by the Tasmanian Native success of their family exists only for the Hen, cannot be dismissed. Florida Scrub Jay ( Woolfenden 1975), the ALLOPARENTAL CARE IN GALLINULES 389

Superb Blue Wren (Malurus cyaneus; Rowley chicks were last observed being fed by a ju- 1965), and the Kookaburra (Dacelo gigas; venile when they were 49-54 days old. Parry 1973). Juveniles and non-parent indi- Family 2 had two broods during this study. viduals in Purple Gallinule families that care When the two offspring from the first brood for the chicks provide a substantial amount were at most 69 days old they were seen feed- of food. Thus, the amount of feeding effort ing the chicks of the second brood. When expended by breeders with helpers is poten- the feeding began, the offspring from the first tially less than that which would be expended brood had juvenile plumage. by breeders lacking helpers. This saving in In addition to feeding chicks, juveniles also feeding effort could be used for other activ- helped the supposed parents defend the fam- ities, such as production of an additional ilys’ territory and the chicks from intruders. brood. Such a benefit was found in the Forty-four percent of the defensive actions Kookaburra (Parry 1973) in which groups were against conspecific gallinules, the re- with helpers managed to nest twice during mainder against other birds, mostly Jacanas a breeding season, although simple pairs did (Jacam spinosa) and Common Gallinules not. In a territory with abundant food, one ( Gallinula chloropus) . pair of Purple Gallinule breeders with helpers Circumstantial evidence indicates that the might produce two or three broods of chicks juvenile helpers in Families 1 and 3 were the in a 5-month period if the amount of time for older siblings of the chicks. In Family 2, egg formation and incubation were short which had four or more adults when obser- enough. At CATIE, a clutch of 3 eggs (Fam- vations began, the filial relationships between ily 6) hatched 22 days after egg-laying com- members of the family are less clear. menced. The incubation period was 20 days Observations suggest that the Purple Gal- for the first egg laid and 21 days for the linules at Turrialba are monogamous, but other two eggs, I-2 days less than that esti- another type of mating system may also exist. mated for clutches in Panama (Gross and van Purple Gallinule helpers may increase the re- Tyne 1929). Thus, the same pair might pro- productive success of their family by reduc- duce successive broods SO-60 days apart, as ing the amount of effort expended by the hypothesized for Family 2. breeders to care for the chicks. This poten- The larger number of broods produced in tial savings in expended energy may be used Family 2 as compared to the other families for production of an additional brood. during this study indicates that helpers may Since the Purple Gallinule is the eighth increase reproductive success. I do not know species of rail for which helpers have been whether any egg or chick predation on broods documented, the habit may be widespread with two chicks occurred before I found in this family. them, but no chicks were lost in any of the observed families after they were first seen. In ACKNOWLEDGMENTS addition, both nests found with three eggs subsequently produced three chicks which I thank the Centro Agronomic0 Tropical de Investi- survived at least to the juvenile stage. Thus, gacion y Ensenanza, Turrialba, Costa Rica whose while additional data are needed, my findings cooperation made this study possible. Special thanks go to R. Morales and G. Budowski of the Depart- support the view that helpers may increase ment of Forest Sciences at CATIE, for their help on breeding success by increasing the number numerous occasions. I am also indebted to J. Leon of broods produced per year in addition to of the Genetic Resources Program for identifying any influence exerted on the survival rates certain plants in the Turrialba pond. of chicks. I am grateful to J. Krekorian who assisted in many ways during this study, and to L. Reeves, who assisted in the field work. Gerald Collier and George SUMMARY Cox provided constructive criticisms of the manu- script and in addition, Collier provided a love for Six Purple Gallinule families (Gallinula mar- birds that has been infectious. Finally, I wish to tinica) were studied for a total of 238 h at thank the American Philosophical Society for part of Turrialba, Costa Rica, from September 1976 the funds that supported this study. through February 1977. Alloparental feeding occurred in four families that had juveniles LITERATURE CITED and chicks. On the average, each juvenile in AILZEHICAN ORNITHOLOGISTS ’ UNION. 1957. Check- families studied in detail fed the chicks about list of North American birds. Fifth ed. Am. as frequently as did each adult. The chicks in Ornithol. Union, Baltimore. BROOKE, R. K. 1975. Cooperative breeding, duet- Families 1 and 2 were fed by the juveniles ting, allopreening and swimming in the Black and three or more adults in each family, These Crake. Ostrich 46:190-191. 390 CHARLES ONEIL’ KREKORIAN

BROWN, J. L. 1972. Communal feeding of nest- Hen, T&onyx motiierii. II. The individual, the lings in the Mexican Jay (Aphelocoma ultra- groun. and the uooulation. CSIRO Wildl. Res. mu&za) : interflock comparisons. Anim. Behav. i7:5^3-90.‘ - - 20:395-403. RIDPATH, M. G. 1972b. The Tasmanian Native CRAIG, J. L. 1975. Co-operative breeding of Pukeko. Hen, T&onyx mortierii. I. Patterns of behaviour. Emu 74:308. CSIRO Wild]. Res. 17:1-51. FRY, C. H. 1972. The social organization of bee- ROWLEY, I. 1965. The life history of the Superb eaters (Meropidae) and co-operative breeding Blue Wren, Malurus cyaneus. Emu 64:251-297. in hot-climate birds. Ibis 114:1-14. ROWLEY, I. 1976. Co-operative breeding in Aus- GREY OF FALLODON, VISCOUNT. 1927. The charm tralian birds. Proc. XVI Int. Omithol. Congr. of birds. Frederick A. Stokes, New York. ( 1974 ) :657-666. GRIMES, L. G. 1976. Co-operative breeding in Afri- RUTHKE, P. 1939. Beobachtungen am Blasshuhn can birds. Proc. XVI Int. Ornithol. Congr. (Fulica atra L.). Omithol. Monatsber. 47:141- ( 1974) :667-673. 147. GROSS, A. 0. AND J. VAN TYNE. 1929. The Purple SKUTCH, A. F. 1961. Helpers among birds. Condor Gallinule (Ionomis martinicus ) of Barro Colo- 63: 198-226. rado Island, Canal Zone. Auk. 46:431-446. TRAUTMAN, B. M. AND S. J. GLINES. 1964. A nest- HAMILTON, W. D. 1964. The genetical evolution ing of the Purple Gallinule (Porphydu mar- of social behavior, 1 and 2. J. Theor. Biol. 7: tinica) in Ohio. Auk 81:224-226. l-16, 17-52. V~~RENCA&, S. L. 1977. Relative fecundity and HAHRISON, C. J. 0. 1969. Helpers at the nest in parental effort in communally nesting anis, Australian passerine birds. Emu 69:30-40. Crotophnga sulcirostris. Science 197:403-405. JENNI, D. A. AND G. COLLIER. 1972. Polyandry WET~LIORE, A. 1965. The birds of the republic of in the American Jacana (Jucana spinosa). Auk Panama. Part 1. Smithson. Misc. Collect. 150: 89:743-765. l-467. LIGON, J. D. 1970. Behavior and breeding biology WILSON, E. 0. 1975. Sociobiology: The new syn- of the Red-cockaded Woodpecker. Auk 87:255- thesis. Belknap Press, Cambridge, Massachusetts. 278. WOOLFENDEN, G. E. 1975. Florida Scrub Jay help- MCILHEKNY, E. A. 1934. Bird city. Christopher ers at the nest. Auk 92:1-15. Publishing House, Boston. WOOLFENDEN, G. E. 1976. Co-operative breeding MEANLEY, B. 1963. Pre-nesting activity of the Pur- in American birds. Proc. XVI Int. Ornithol. ple Gallinule near Savannah, Georgia. Auk Congr. ( 1974) :674-684. 80: 545-547. ZAHAVI, A. 1974. Communal nesting by the Ara- MEISE, W. 1934. Zur Briitbiologie der Ralle Lateral- bian Babbler, a case of individual selection. Ibis Zus Zeucopyrrhus (Vieill.). J. Ornithol. 82:257- 116:84-87. 268. ZAHAVI, A. 1976. Co-operative breeding in Eu- PARRY, V. 1970. Kookaburras. Lansdowne Press rasian birds. Proc. XVI Int. Omithol. Congr. Pty Ltd., Melbourne. ( 1974) :685-693. PARRY, V. 1973. The auxiliary social system and its effect on territory and breeding in Kooka- Department of Zoology, San Diego State University, burras. Emu 73:81-100. San Diego, California 92182. Accepted for publica- RIDPATH, M. G. 1972a. The Tasmanian Native tion 3 October 1977.