Revista Chilena de Historia Natural 69: 343-349, 1996 Frugivory and seed dispersal of Podocarpus nubigena in Chiloe, Chile

Frugivorfa y dispersion de semillas de Podocarpus nubigena en Chiloe, Chile

1 2 MARY F. WILLSON , CARLOS SABAG , JAVIER FIGUEROA2 and JUAN J. ARMEST02

1Forestry Sciences Laboratory. 2770 Sherwood Lane, Juneau AK 99801, USA. E-mail: /s =M.WILLSON/OUI = [email protected] 2 Departamento de Biologia, Facultad de Ciencias, Universidad de Chile, Casilla 653, Santiago, Chile. E-mail: [email protected]

ABSTRACT

We studied seed dispersal of mañío macho (Podocarpus nubigena Lindl.) by birds in rainforests of Chiloe. Austral thrushes (Turdusfalcklandii) were the predominant agent of seed dispersal, carrying about 18% of seeds at least three tree-crowns away from the parent tree. Several other passerines also dispersed some seeds. Chilean pigeons ( Columba araucana) virtually always dropped seeds below the parent. Although seeds did not germinate in the laboratory, in field experiments, seeds germinated regularly. The density of seedlings and juveniles exceeded 151m 2 under female mafiios and /m2 under other trees in the forest. Key words: frugivory. seed dispersal, fleshy fruit. Chilean rainforest.

RESUMEN

AI estudiar la dispersion de semillas por !as aves en el bosque lluvioso de Chiloe se ha encontrado que Ios zorzales (Turdus falcklandii) son Ios principales agentes dispersantes del mañío macho (Podocarpus nubigena Lindl.), transportando alrededor del 18% de !as semillas a! menos a una distancia equivalente a dos copas de arboles de la planta madre. Además del zorzal, otros paseriformes tambien dispersan algunas semillas. Las torcazas (Columba araucana) casi siempre dejan caer !as semillas debajo de la planta madre. Aunque no se obtuvo germinacion en laboratorio ni en experimentos de campo. !as semillas germinaron regulannente en la naturaleza. La densidad de plantulas y de juveniles excede 151m2 debajo de Ios árboles hembras y /m2 debajo de otros arboles dentro del bosque. Palabras clave: Frugivoria, dispersion de semillas. frutos carnosos, bosque lluvioso de Chile.

INTRODUCTION only for understanding fundamental species interactions that help determine community The temperate rainforest of southern Chile is organization and function (Will son 1991 ), characterized by a high diversity of fleshy- but also for practical applications, such as fruited trees (Armesto & Rozzi 1989). An maintaining natural regeneration of species important fleshy-fmited canopy species of in forests that are increasingly fragmented. montane forests between 40° and 42°S and Here we document seed dispersal ecology of lowland forests south of 42° S is the maf\fo, providing information on avian frugi- dioecious gymnosperm "maf\fo macho" vores, seed-handling behavior and location (henceforth "maf\fo; Podocarpus nubigena of seed deposition, seed predation, seed ger- Lindl.; Podocarpaceae)., Although many mination, and seedling distributions. species of forest birds eat fmits and disperse the seeds of rainforest plants in southern Chile (Armesto et al. 1987, Sabag 1993), STUDY SITE AND METHODS seed-dispersal ecology has not been studied for any species in these forests. The study of We use the Chilean common names of the seed-dispersal ecology has importance not birds in the text, by preference. English

(Received 20 July 1995: accepted 16 May·J996: managed by Guillerrno Riveros) 344 WILLSON ET AL. common names and scientific names are The "fruits" of mafifo are unlike those of given in Table 1. some other gymnosperms and most verte- The study site was located in the north- brate-dispersed angiosperms, in that the east part of Isla Grande de Chiloe, just west edible portion does not enclose the seeds but of Linao. An extensive lowland forest is rather consists of an aril at the base of the comprised principally of tepual, a swamp seed. The aril is red, sweet to human taste forest in which tepu (Tepualia stipularis (about 32% sugar, 1% lipid, 4% protein; (H. et A.) Griseb.), a small tree often with analyses by the Palmer Research Center, creeping stems, is common. Coigiies (Notho- Palmer, Alaska), and juicy (about 81% fagus dombeyi (Mirb.) Oerst. were common water) when ripe. The mature seed is black. emergents, and mafifo was frequent in the The aril is usually about the same size (and canopy and as juveniles in the understory. often shape) as the seed: the aril averages 8.5 Selective logging (especially of coigiie and mm in length (SE = 0.7) and 8.2 mm in mafifo) had thinned the edge of this forest for diameter (SE = 0.9), the seed averages 9.7 a distance of about 100 m. All focal trees for mm in length (SE = 0.6) and 7.9 mm in avian observations were located within or diameter (SE= 0.6) (Figure 1). The shape of near this disturbed edge, where visibility was relatively good. Maiifo produces ripe fruits in spring, at a time of the year when few SEED (BLACK) ARIL (RED) other fruits are available to frugivores. Most vertebrate-dispersed plants produce mature fruits in summer and autumn. One or two observers watched each group of focal trees for 2-3 hours per morning, recording species of birds foraging in ma- iifos, seed-handling behavior and fate of seed, and other behavior of frugivores in- cluding flight distances and aggression. In all, there were 95 hours of observation in early October 1993, largely before the arrival of the migratory, frugivorous ffo-ffo, and 93 1cm hours in late October 1994, after the ffo-ffos had arrived. Foraging birds were often Fig. 1: A mature dispersal unit of Podocarpus nubigena, consisting of seed and aril. Drawn ap- hidden behind the trunk or in dense foliage, proximately 3x actual size. so it was seldom possible to record all seeds Unidad de dispersi6n madura de Podocarpus nubigena, con- handled by the birds. Over 25 maiifo trees sistente en una semilla y un arilo. El dibujo es approxima- were included in the observations. damente 3 veces el tamaiio real.

TABLE I

Fruit-eating avian visitors to mafifo trees: percent of seeds (of observed fate) handled by each species. + =< 1% Aves frugfvoras que visitaron Ios arboles de maiifo macho: porcentaje de semillas (con destino observado) manipulados por cada especie. + = < I%

Scientific name Chilean common English common Early Oct., 1993 Late Oct., 1994 name name N = 456 N =955

Turdus falcklandii zorzal austral thrush 91% 60% Curaeus curaeus tordo austral blackbird 6 + Phrygilus patagonicus cometocino Patagonian sierra-finch 3 1 Elaenia a/biceps ffo-ffo white-crested elaenia + 2 Pyrope pyrope diuc6n fire-eyed diucon + Aphrastura spinicauda rayadito thorn-tailed rayadito + Calumba araucana torcaza Chilean pigeon 36 Enicognathus leptorhynchus choroy slender-billed parakeet 1 FRUGIVORY AND SEED DISPERSAL OF PODOCARPUS NUB/GENA 345

both seed and aril is usually oval or barrel- RESULTS shaped, and the mature aril does not resemble the illustrations in Rodrfguez et al. The zorzal was the most frequently recorded (1983) or Hoffman (1982). One seed per aril consumer of mafifo in both field seasons, and is common, but occasionally two seeds are a number of other birds harvested "fruits" borne on a single aril, which is generally (Table 1). Although the ffo-ffo is a major larger and less symmetrical than usual. frugivore in these forests (Sabag 1993), it ac- counted for only a small proportion of seeds Freshly fallen seeds collected from the handled in this study. forest floor in October 1993 were pre-treated ° Zorzales handled seeds and arils in several by holding at 4 C for 60 days. They were ways. They often swallowed the entire dis- then placed in a germination chamber at the persal unit (seed + aril), regurgitating the Facultad de Ciencias, Universidad de Chile, seed several minutes later. After swallowing Santiago, for ten months at 20° C with 12 several dispersal units, a zorzal usually sat hours of exposure to light each day. Sample quietly for several minutes, and then ejected sizes were 100 seeds without arils, distribut- one to several seeds. Occasionally a seed was ed equally among ten germination dishes, brought up and reswallowed. Seed ejection and 100 seeds with attached arils, similarly was usually accomplished with little overt distributed. In the field 200 seeds with arils physical effort, and close observation was and 200 without arils were placed on the needed to see it happen. Alternatively, zorza- ground in the study site, in wire-mesh ex- les frequently held the dispersal unit by the closures in October, 1993. The exclosures aril and struck it against a branch, often were pressed into the ground about 2 cm many times, until the seed (usually) fell off and held in place with pegs, to prevent ready or the whole dispersal unit was dropped. Oc- access by ground-foraging vertebrates. There casionally, however, the seed remained at- were 20 seeds/exclosure and 10 exclosures tached and the whole unit was then swallow- per treatment (with arils or without arils). ed. Finally, zorzales sometimes carried The ex closures were set out in pairs (one dispersal units away in their bills. The relative with arillate seeds and one with arilless frequency of swallowing and striking differ- I seeds) along a 00 m transect. Seeds in the ed between seasons (Table 2). Dispersal units exclosures were monitored for germination bearing two seeds were uncommon; they in January and November, 1994. appeared to be more difficult to handle than Seed predation was examined experi- one-seeded units and were often dropped. mentally in October 1993. Ten seeds without Zorzales were commonly very aggressive arils and ten arillate seeds were placed on while foraging on mafifo. They chased each the ground at each of twelve stations and other, sometimes vigorously, on most oc- checked the following day for evidence of casions when more than one individual was seed removal or damage. The experiment present. However, on some occasions, sever- was repeated nine times during October. al zorzales foraged simultaneously in the Seedling and juvenile density, as a function same tree, so they cannot be said to be suc- of distance from female trees, was surveyed cessfully territorial in mafifos. Zorzales also in October 1993. Two 1-m2 quadrats were sometimes chased ffo-ffos and cometocinos haphazardly placed under each tree sampled. but were not observed to try to displace tor- Five female individuals of mafifo, 13 other cazas or tordos. kinds of trees (male individuals of mafifo, Zorzales that foraged in the focal trees coigiies, canelos [Drimys winteri J. R. et G. often arrived from and departed to considera- Forster]) adjacent to females, and five trees ble distances. We often observed them flying at least three tree-crowns away from fe- over a large, shrubby field adjacent to the males ("far") were sampled. Differences in forest, a minimum distance of several hun- average densities of young maiifos at three dred meters from the next nearest stands of distances (under female, under adjacent trees. Furthermore, we found several mafilo tree, or "far") were tested by Kruskal-Wall is seeds (and seedlings) in this field, over 100 test. m from the forest edge. 346 WILLSON ET AL.

TABLE 2

Seed-handling behavior of frugivores (those with> 10 observations). N = number of seeds. The data are percentages of occurrence, for each bird species Comportamiento de manipulacion de semillas porI as aves frugfvoras (solo aquellas especies con mas de I 0 observaciones). N = numeros de semillas. Los datos son porcentajes de occurrencia, para cada especie de ave

Species Year N (seeds) Swallow Strike Drop Carry

zorzal 1993 414 19% 64% 3% 14% 1994 575 61 29 9 I cometocino 1993 14 92 8 1994 12 75 8 17 ffo-ffo 1994 16 6 75 19 torcaza 1994 339 98 2 tordo 1993 27 4 59 30 7

Torcazas were only observed on seven Few seeds disappeared from the experi- occasions, but they sometimes stayed a long mental seed-predation plots, but slightly more time (almost an hour) in the same tree, forag- arillate seeds (5%) were taken than seeds ing almost continually. They customarily without arils (2%); removal of arillate seeds plucked the dispersal unit, bit off the seed, exceeded that of arilless seeds on 7 of 9 days which fell to the ground, and then swallowed (Sign Test, P = .09). No seeds were destroy- the aril, but sometimes they picked and ed in situ, although some appeared to have dropped the seed before plucking the aril. In been chewed slightly. On the other hand, either case, they did not serve as effective 8% of the arils were chewed, pecked, or con- dispersal agents. sumed. Cometocinos and tordos usually consumed Experimental germination tests failed the aril while still in the parent-tree crown, completely, both in the germination chamber but occasionally dispersal units were carried and in the field. However, it is clear that na- away (Table 2). The tordo often held a dis- tural germination occurs regularly, because persal unit against a branch with its foot and seedlings and juveniles (up to 1 m tall) are pecked at the aril, or struck off the seed as common. Their average density under female the zorzales did. Cometocinos usually nib- mafiios was 15.5 individuals/m2 (n = 10, SE bled small bites from the aril and dropped the = 2.9). In contrast, the density under adjacent seed. Rayaditos usually pecked at the aril in trees was only 3.4 individuals/m2 (n = 26, SE situ but were once observed to carry off the = 0.5). It was difficult to find sites within the whole dispersal unit. Both ffo-ffos and diucones commonly knocked off the seed before swallowing the aril but sometimes TABLE 3 carried fruits away from the parent tree. Deposition patterns of mafifo seeds Choroyes bit off the aril and let the seed handled by avian frugivores, 1994 data. fall, during our focal-plant observations. Separate entries only when n ~ 10 seeds per However, on the forest floor we sometimes bird species, expressed in percentages found accumulations of seeds that had been Patrones de deposicion de las semillas de mafifo por las aves split neatly open and the embryo extracted. frugfvoras. Datos de 1994. Entradas separadas solo cuando We suspected this to be the work of choroyes n 2: 10 semillas por especie, expresadas en porcentajes (Sabag, pers. obs.). Species N Parent tree Adjacent tree "Far" We estimated that 75% of all seeds were dropped beneath the parent trees in 1994, and zorzal 353 52% 12% 36% cometocino 10 60 10 30 that 18% were carried "far" from the parent. ffo-fio 12 83 17 The remainder (6%) fell under trees adjacent torcaza 339 100 to the parent (Table 3). All birds 730 75 6 18 FRUGIVORY AND SEED DISPERSAL OF PODOCARPUS NUB/GENA 347 study area that were distant from female away from the parental vicinity (Howe et al. mafifos: a small sample (n = 6) yielded an 1985). Other studies also indicate considera- average density of 1.4 (SE = 2.6), only ble variation in the probability of a seed slightly lower than that under trees adjacent travelling far from the parent (e.g., Debus- to females. The densities of young maiifos sche et al. 1985, Dirzo & Domfnguez 1986, differed significantly at different distances Holthuijzen et al. 1987, Debussche & Isen- (Kruskal-Wallis H = 16.6, P < .01). mann 1994). Our estimate that zorzales carry about 18% of seed "far" from the parent thus falls within the observed variation for other DISCUSSION species in other places, but good compari- sons are not possible with present informa- The zorzal, a member of a genus well-known tion (Willson 1993). as frugi vores and seed vectors in North The importance of mafifo arils for zorza- American and Europe (Willson 1986, Snow les is not known. The crop ripens early in the & Snow 1988), was clearly the predominant breeding season (usually October), when agent of seed dispersal in this study. Never- most zorzales are nest-building or incubat- theless, zorzales probably deposited most of ing, but some dispersal units may remain on the seeds they handled immediately under the tree as late as December. Several obser- the parent tree. Most of the other visitors vations suggest that mafifo provides an were rare and/or deposited an even higher important food source at this time. In Octo- proportion of seeds below the parents. High ber, some zorzales fly considerable distances levels of aggression by proprietary zorzales and spend long periods foraging in mafifo probably contributed to the frequency at trees. The level of aggression among zorzales which avian foragers retreated to nearby foraging in mafifos in Chiloe was much trees to consume mafifo "fruits" and also to higher than that of zorzales foraging on the broad distribution of seeds and seedlings. Drimys fruits in Tierra del Fuego (Willson & Furthermore, zorzales appeared to have dif- Sabag, pers. ohs., February 1995). Moreover, ficulty handling two-seeded arils and fre- the availability of other fruits is very low at quently dropped them; failure of dispersal this time of year. Incidentally, on two occa- away from the parent tree may create selec- sions, we also saw zorzales harvesting inner tion pressure against two-seeded dispersal bark of maiifo for nest material. units. Zorzales sometimes travel some dis- Maiifo seed failed to germinate in the tance to visit fruiting maiifos, commonly laboratory, as reported also by Urrutia cross open fields, and are capable of dis- ( 1986). The seeds also did not germinate in persing seeds from mafifos in perturbed and our field tests, although natural recruitment isolated woodlots as well (Willson et al. is excellent in this forest. It is possible that 1994). Thus, maiifos growing in fragmented germination of mafifo seeds requires more forests may still obtain seed-dispersal ser- time or some specific interactions with forest vices from zorzales. soil that was somehow precluded by our Although vertebrate-dispersed seeds may exclosures. In contrast to other trees such as be carried very long distances in some cases some Nothofagus and Drimys (which com- (Fleming & Heithaus 1981, Willson 1993), monly germinate on fallen logs), mafifo ger- few studies assess the seed shadow of wood minates directly on the forest floor. plants in ways that allow direct comparison Although some seeds were removed from with our study. For Virola surinamensis in the seed-predation plots, it is not clear that Panama, the proportion of seeds carried more removal constitutes predation in this case. than an estimated three crown-diameters The arils of arillate seeds were often chewed from the parent plant was 65% for three and sometimes destroyed, in situ, and re- species of large birds (toucans, guan), but moval may have occurred because the < 4% for two smaller bird species (trogon, consumer sought the aril. If so, this would motmot; Ho we & Vande Kerckhove 1981 ), constitute dispersal. The attractive properties although another study of the same species of the aril may explain the slighltly higher suggested that very few seeds were carried removal rates of seeds with arils. The hard 348 WILLSON ET AL.

covering of the embryo is resinous and about 1135 and 195046 I (Armesto) and the 2 mm thick, which may deter some would-be National Science Foundation-International predators. Although there was little good Programs (Willson). We are grateful to evidence of predation in the experiment, Melvin Koenig for permission to use his seeds that fall naturally to the forest floor forest as a study site, and to E. Anderson, L. usually disappear within about two months. Solchenberger, K. Bardon, T. Partner, and H. The most likely predators on the forest floor Puckey for field assistance. We thank A. are probably rodents. Reasons for the dif- Traveset for comments on the manuscript ference between experimental and natural and E. Anderson for Figure 1. This is contri- removal of seeds are not known. bution No 6 from Estaci6n Biol6gica Senda Although zorzales can ingest several Darwin, Ancud, Chiloe, Chile, a field station arillate seeds at one sitting, they often established by the Institute of Ecological remove the seed before ingesting the aril. Research Chiloe. Knocking off the seed commonly required vigorous action (often 30-60 strikes). On the other hand, ingestion of the whole dispersal LITERATURE CITED unit requires internal handling for several minutes and limits the number of arils eaten ARMESTO JJ & R ROZZI (1989) Seed dispersal syndromes in the rain forest of Chiloe: evidence for the impor- per unit time, because the seeds then occupy tance of biotic dispersal in a temperate rain forest. space in the upper digestive tract. We ob- Journal of Biogeography 16: 219-226. served that the relative frequency of inges- ARMESTO JJ, R ROZZJ, P MIRANDA, & C SABAG (1987) Plant/frugivore interactions in South Ameri- tion vs knocking off the seed varied among can temperate forests. Revista Chilena de Historia birds and/or among trees, suggesting that the Natural60: 321-336. costs of aril harvesting may also vary, with DEBUSSCHE M, P ISENMANN (1994) Bird-dispersed seed rain and seedling establishment in patchy possible consequences for both foraging and Mediteranean vegetation. Oikos 69: 414-426. seed dispersal. DEBUSSCHE M. J LEPART, & J MOLINA (1985) La dis- Zorzales are probably more common in semination des plantes a fruits charnus par les oiseaux: role de la structure de la vegetation et impact sur la this part of Chiloe now, as the forest is in- succession en region mediterraneenne. Acta Oecol- creasingly cleared for timber-harvest and gica/Oecologia Generalis 6: 65-80. agriculture (Will son et al. 1994 ), than in the DIRZO R & CA DOMINGUEZ (1986) Seed shadows, seed predation and the advantages of dispersal. In Estrada time of Darwin over 150 years ago (Willson A & Fleming TH (eds) Frugivores and seed dispersal: & Armesto, 1996). This frugivore would 237-249. Junk, Dordrecht, Netherlands. probably be less common in primeval forests FLEMING TH & ER HEITHAUS ( 1981) Frugivorous bats, seed shadows, and the structure of tropical forests. than in the present landscape of woodlots Biotropica 13 (Suppl.): 45-53. and fencerows, but it is not clear what (or if) FLOYD AG ( 1989) Rainforest Trees of Mainland South- other frugivores might have been the princi- eastern Australia. lnkata, Melbourne. 420 pp. HOFFMANN AE (1982) Flora Silvestre de Chile: Zona pal aboriginal dispersal agents of mafifo. Araucana. Ediciones Fundaci6n Claudio Gay, Santia- Comparison of frugivory and seed disper- go, Chile. 258 pp. sal with other species of Podocarpus in Chile HOL THUIJZEN AMA & TL SHARIK (1985) The red cedar (luniperus virginiana L.) seed shadow along a fen- and elsewhere in the southern hemisphere celine. American Midland Naturalist 113: 200-202. would be interesting. Seed size and color, HOL THUIJZEN AMA. TL SHARIK, & JD FRASER ( 1987) aril size and color, growth form, and habitat Dispersal of eastern red cedar ( Juniperus virginiana) into pastures: an overview. Canadian Journal of Bo- all vary (Salmon 1980, Floyd 1989), as do tany 65: 1092-1095. the potential disperser assemblages. These HOWE HF & GA VANDE KERCKHOVE (1981) Removal systems might be suitable for determination of wild nutmeg (Virola surinamensis) crops by birds. Ecology 62: 1093-1106. of the relative importance of phylogeny vs. HOWE HF, EW SCHUPP, & LC WESTLEY (1985) Early regional ecology in the evolution of fruiting consequences of seed dispersal for a neotropical tree characteristics. (Vim/a surinamensis). Ecology 66: 781-791. RODRIGUEZ R, 0 MATTHEI, & M QUEZADA (1983) Flora Arb6rea de Chile. Universidad de Concepci6n, ACKNOWLEDGMENTS Chile. 408 pp. SA BAG C. ( 1993) El rol de !as aves en la dispersion de semillas en un bosque templado secundario de Chiloe This project was a small part of larger studies (42° S). MS thesis, Facultad de Ciencias, Universidad funded principally by Conicyt, grants 92- de Chile. Santiago. 79 pp. FRUGIVORY AND SEED DISPERSAL OF PODOCARPUS NUB/GENA 349

SALMON JT ( 1980) The Native Trees of New Zealand. Hei- WILLSON MF (1993) Dispersal of seeds by frugivorous nemann. Reed, Auckland. 384 pp. animals in temperate forests. Revista Chilena de His- SNOW B & D SNOW (1988) Birds and Berries. Poyser, toria Natural 64: 537-554. Calton, U. K. 268 pp. WILLSON MF (1993) Dispersal mode, seed shadows, and URRUTIA JR (1986) Analisis biblignifico y pict6rico de colonization patterns. Vegetatio I 07/108: 261-280. semillas y sus procesos germinativos para 32 espe- WILLSON MF, TL DE SANTO, C SABAG & JJ ARMES- cies forestales nativas. Tesis Ing. Forestal, Facultad TO ( 1994) A vi an communities of fragmented south- de Ciencias Forcstales, Universidad Austral de Chile, temperate rainforests in Chile. Conservation Biology Valdivia. 150 pp. 8: 508-520. WILLSON MF (1986) Avian frugivory and seed dispersal WILLS ON MF & JJ ARMESTO ( 1996). The natural history in eastern North America. Current Ornithology 3: of Chiloe: on Darwin's trail. Revista Chilena de His- 223-279. toria Natural 69: 149-161.