Do precipitation and food affect the reproduction of brown brocket deer Mazama gouazoubira (G. Fischer 1814) in conditions of semi-captivity?

The births of brown brocket deer (Mazama gouazoubira) in a secondary lower montane forest called “yunga” in northwestern Argentina were compared with rainfall. Analyses were performed with rainfall and flower-fruit fall in an attempt to determine the possible importance of these seasonal variables in birthing. The births were not directly correlated with rainfall, but rather with the flower and fruit fall of exotic plant species. This may be related to favor the development of fawns, which eat the new and more digestible plant parts, accessible one month after their births. The non-seasonal births observed around the year could be related to the selection by the deer of some plant species that have been introduced into the region (Prunus, Morus and Psidium), have a longer fruiting span than the scarce native plant species.


INTRODUCTION
The brocket deer species (Mazama spp.) constitute a diversified Neotropical genus with six to eleven species, some of them of recent discovery or classification (Duarte and Merino 1997, Geist 1998, Medellín et al. 1998, Putman 1988, Weber and González 2003).The taxonomic value of the genus is also in doubt, since it could be polyphyletic and in fact a species, M. americana, is closer to the genus Odocoileus than to other Mazama species (Gilbert et al. 2006).Despite the broad distribution and apparent abundance of some species of brocket deer, most of their biology, including their reproductive cycle, remains obscure, although the latter is considered to be not seasonal since it lives in the tropics, with few seasonal climatic changes (Barrette 1987).by anti-predator strategies and food availability, s al patterns in climate tend to be one of the main accounting for the variance in ruminant birthlength.In any case, according to Rutberg (1987) t and subtropical species of deer may give birth at an of the year, showing only weak breeding peaks.
Some details are known about the reprod of the red brocket deer Mazama americana (Er 1777).This species reproduces along the year, al there seems to be a more marked concentration o in the months considered to be most favorable development of fawns (Branan and Marchinton Thus, in Venezuela, fawns are most frequently ob during the periods of the year with the lowest r (Bisbal 1994).In Peru, Gardner (1971) describ "main" -2010/8/6 -18:55 -page 630 -#2

630
JUAN P. JULIÁ and SALVADOR J. PERIS brocket deer is also relatively seasonal in Surinam, where births are concentrated between September and April, in the first -and more moderate -of the two periods of annual rains in the region (Branan and Marchinton 1987).In the Misiones province (Argentina), red brockets reproduce with greatest frequency between August and October (Crespo 1982).In contracts, in northeastern Peru, the species did not show reproductive seasonality (Hurtado- González and Bodmer 2006).Macnamara and Eldridge (1987) observed that in captivity the red brocket reproduces at the same rate through the year.
Less is known about the reproductive biology of the brown brocket M. gouazoubira (G.Fischer 1814), (Richard and Juliá 2001a).However, in wild may show birth peaks (Stallings 1986, Frädrich 1987, Bisbal 1994).Also, as M. gouazoubira is a polyphletic species (Weber and González 2003, Hassanin pers. com.), the different papers on its reproduction could also refer to different species.
Few data are available about the reproduction of the deer in subtropical forests, such as those located at the southern limits of its distribution, like the premontane forest in northwest Argentina.This eco-region has a seasonal climate, with a dry season in which the decidous trees lose their leaves (Prado 1995), and these variations could influence the reproduction strategy of the deer (Richard et al. 1995a).Although brocket deer forage on about 75 plant species, they prefer only a few of them (Richard and Juliá 2001b).Since the reproductive parts of plants (flowers, fruits and seeds) are an important, although seasonal, resource in the diet of brown brocket deer, with up to 68% of the diet (Gayot et al. 2004), we tested for possible correlations between the availability of the most important flowers and fruits in the deer diet and the births of fawns.Currently, these select items come from four exotic species that are very palatable to the deer and over the last 50 years became widely distributed on the lower slopes of the San Javier hills: namely, peach (Prunus persica), mulberry (Morus alba and M. nigra) and guava (Psidium guajava).The present study analyses the reproductive data of the spe-

MATERIALS AND METHODS
The study was carried out at the Reserva Experimental Horco Molle (REHM), a protected area supervised by the Faculty of Natural Sciences and the Miguel Lillo Institute of the National University of Tucumán, Argentina (Richard 2000).It is located on the eastern slopes of the San Javier hills (26 • 38 -26 • 57 S to 65 • 26 -65 • 20 W) in the province of Tucumán (northwest Argentina).Mean annual rainfall in the zone is about 1313 mm.The study area has a dry season, with less than 75 mm/month on average, that extends from May to October, and a rainy and hot season between November and April, with more than 95 mm/month of average rainfall.The original premontane forest vegetation has been strongly altered, and the vegetation now comprises pastures, shrub and secondary forests (Aceñolaza 1989).In 1988, six brown brocket deer were confined in an enclosure of around 18 ha with the same natural vegetation as on the lower slopes of the San Javier hills, located near the study area.The species is common in the REHM outside the enclosure.The origin of the research individuals come from the surroundings hills (2) and the rest of the herd, from a Chaco forest of Santiago del Estero province.Both, Chaco and premontane forests show seasonality, although the last is a drier region.In spite of the long study period, the enclosure, although large enough (>25 ha) and with habitat similar to its surroundings (including vegetation and occasional predation by ocelot (Felis pardalis), the mixed stock origin of the animals could affect some of the data interpretation.The number of deer in the area fluctuated from eight to 16 individuals during the 14 years of the study, with an average between 10 to 14 animals (on average there were only 2-3 males at the same time).Data were obtained daily by wardens and the authors from September 1988 to June 2001, and sampling was performed throughout the enclosure, following marked routes.Sampling units were the direct observation of births, fawns, or the mothers with their fawns.Three miscarriages were observed (Juliá and Richard 2001), and their estimated dates of birth were added to the data.
could not be determined, such observations were discarded.Births were ordered by month and possible conceptions were estimated on the basis of the period of gestation of the species, which is around 7 months (Barrette 1987).
The only seasonal variables considered were rainfall and the availability of palatable fruits for deer and the fall of flowers (mulberry) (Richard and Juliá 2001b).
The importance of the plants as food was calculated from 1991 to 1995, following direct observations 4-6 hours per day and 3-5 days per week.Deer were observed at a distance of 3 to 10 m, which permits identification of the plants and their parts consumed, and the proportion of fruits and other plant pants in diet were estimated by direct observation (Richard and Juliá 2001b).The observations were established by combining the presence/absence of fallen flowers/fruits under trees from a sample of 6 peach, 10 guava and 10 mulberry trees and the percentage of these fruits in the diet of the deer.We established an estimator that combined the presence of food -the proportion of samples (n = 840) with fallen fruit or flowers -and its importance in the diet by ranks.These were: 0 = almost complete absence of fallen flowers and fruits and importance in the diet of less than 1%; 1 = low availability flowers and fruits (10/15% of the sample and importance in the diet from 1-10%); 2 = medium availability flowers and fruits (more than 20% and less than 50% of the samples) and importance in the diet between 11 and 20%; 3 = medium-high availability flowers and fruits (51-100% of the samples) and importance in the diet greater than 20%.
Monthly rainfall data, obtained from a local station, were from 1978 onwards (Ortiz pers.comm).A Mann-Whitney U -test was performed to determine whether there were significant differences between the number of births per month in the dry and rainy seasons.The Spearman correlation test evaluated the relationship between rainfall and births and between the presence and consumption of flowers/fruits and fawn births.

RESULTS
significant difference was observed in the birth r tween the dry and rainy seasons (Mann-Whitney = 8.5), and although 65% of the births appeared concentrated within the rainy season (Fig. 1), no s cal correlation was obtained between births and r (r s = 0.407, d.f.= 12).However, it is impor point a significant correlation observed between erage rainfall and the births of the previous month 0.711 ≤ 0.02, d.f.= 12) (Fig. 1).Also, the sta relationship between the births of fawns and th sumption by deer of the plant reproductive parts ( fruits) was highly significant (r s = 0.706 ≤ 0.01 12) (Fig. 2).The fawns began to eat small amo solid matter around the first week of life, but th not begin to forage with their mothers until they a month old.The months with the greatest availab exotic fallen fruits were September, October, N ber, January, March and April.

DISCUSSION
According to the data obtained here, the brockets to reproduce with some slight peaks in the year, ported for Venezuela (Bisbal 1994) and Bolivian (Noss et al. 2003), but in contrast with the Pe Amazon (Hurtado-Gonzáles and Bodmer 2006).have been observed throughout the year in habita a certain degree of climatic seasonality, such Paraguayan Chaco (Stallings 1986) and brocket captivity in temperate regions (Frädrich 1987).way, it is important to keep in mind that if M. zoubira is a poliphyletic species, those data co fer to different species.In any case, the length breeding season in most mammal species, from r to elephants, is affected by population density, w extended period at low densities (Delany 1982) density in the study area was very high in comp with the maximum natural population, estimated to 12.55 brockets/km 2 (Ayala and Noss 2000).It ficult to establish whether population density infl the breeding season of brocket deer because oth "main" -2010/8/6 -18:55 -page 632 -#4 The dispersion of deer births over the year could be related in tropics with variability in rainfall between years, with peaks of births in those seasons with relatively better climatic conditions (Branan andMarchinton 1987, Bisbal 1994).However, no statistical analysis was performed by the last authors to sustain their comments.Slight birth peaks were observed in the REHM, although these could not be directly correlated statis-between the abundance and availability of reproductive parts of the plants and births (Fig. 2).In other words, rain and fructification influence the reproduction seasonality of brown brocket deer in the study area.
When available, the reproductive parts of plants (fruits, seeds, and to a lesser extent flowers) are some of the main components in the diet of brocket deer in Peru (Bodmer 1989), Surinam (Branan et al. 1985), tant fruits in the diet of the brocket (Morus spp), but also with the beginning of the period of greatest availability of young green shoots (Richard et al. 1995a).According to Demment and Van Soest (1985), small ungulates such as Mazama have poor efficiency in the digestion of fruits and mainly derive their nutritional uptake from seeds (Bodmer 1989).The birth of fawns in the rainy season could be advantageous for the mother, since they would have a greater quantity and quality of food (flowers, fruits and young green parts), appropriate for lactation.Additionally, the availability of forage for the fawns would accelerate their weaning.
Fawns are born one month before the beginning of the periods of greatest availability of fruit, shoots, and young leaves, thus fawns have a readily digestible and available diet one month later, when they begin to feed by themselves.In spite of the exotic nature of the fruit trees which have mostly replaced the natural fruitbearing species in the premontane forest, they fructify in a period similar to that of autochthonous trees.Indeed, indigenous plant species such as Allophylus edulis, Xylosma pubescens, Carica quercifolia, Eugenia pungens, Blepharocalyx gigantea, Cupania vernalis and Phoebe porphyria bear fruits from November to March (Digilio and Legname 1966), coinciding with the rainy season.The fact that 65% of the births were observed in the rainy season -although not statistically significant -may indicate the natural disposition of the deer to give birth coinciding with the sprouting of native plants.The relatively low availability of these plant species on the low slopes of the San Javier hills, given their strongly modified vegetation over the last 50 years, may hinder the detection of any relationship between flowers and fruits from native plants and deer births.In fact, the longer fructification around the year of exotic plants (from September to April -8 months) as compared with that of native plants, just 5 months (Boletta et al. 1995), underscores the adaptability of the deer to exploit these new and (owing to their availability) more profitable resources.The longer fruiting span in exotic plants could be related to the lengthened

ACKNOWLEDGMENTS
We wish to thank the R.E.H.M. and its former tor, E. Richard, for offering access to files and lations, as well as C.P. Juárez, K. García, K.So Juárez, M.C.Cocimano, R. Delgado, and I. Palac their collaboration.J.P. Juliá acknowledges a S AECI grant.An early draft was improved by th ments of A. Loison and A. Zillikens and three mous reviewers.

Fig. 1 -Fig. 2 -
Fig. 1 -Monthly precipitation compared with births in the previous month of M. gouazoubira in the Horco Molle Experimental Reserve-REHM from September 1988 to June 2001.