Revista Chilena de Historia Natural 60: 276-283, 1987 Shrub defoliation in the Chilean matorral: what is its significance? Defoliaci6n de arbustos en el matorral chileno: ¿cuál es su significado? EDUARDO R. FUENTES, ALDO POAINI & JUAN D. MOLINA Facultad de Ciencias Biologicas, Pontificia Universidad Catolica de Chile. Casilla 114-D. Santiago - Chile. ABSTRACT This paper reviews the available information on shrub defoliation in the Chilean matorral. The phenomenon and its patterns in space and time are described. The roles of climate, predators, plant defenses and competition are analyzed within a comprehensive model. Key words: Defoliation, climatic constraint, predation, competition, plant defenses. RESUMEN Este articulo recopila la informacion disponible acerca del papel de Ia defoliacion en el matorral chileno. Se describe el fenomeno y sus manifestaciones en el tiempo y en el espacio. Luego se discute un modelo en que se destacan Ia im- portancia del clima, depredadores, defensas de las plantas, y de la competicion. Palabras claves: Defoliacion, restricciones climaticas, depredaci6n, competici6n, defensas de las plantas. INTRODUCTION 1979, Fuentes & Etchégaray 1981) have This contribution is a complement to E. shown that herbivores can greatly modify Bucher's presentation (in this volume) on the physiognomy of the landscapes of the role of native grazers and browsers in central Chile. We do not intend to further South America. Similarly to what seems to discuss these phenomena. be the Argentinian case, in the Chilean Instead we will identify potentially in­ temperate areas there are still not many teresting questions in relation with a still studies of this type. Most studies of inter- little understood phenomenon: the defoli- actions between terrestrial native herbi- ation of large woody plants in the mediter- vores and plants have been done rather ranean-type shrublands. Our intention is to recently and mostly by people associated provide a general scheme where present to our research group. For that reason ci- observations and future experiments could tations made will look somewhat restricted, fit, rather than a well proven model with but they only reflect the youth of the field "definite answers". As it will be seen, we are in Chile. still far from those answers. Rather than attempting to recapitulate Although physiognomically similar all that has been done, we will focus only formations exist in California and the Me- on one question and will attempt to explain diterranean Basin, as well as in South Afri- its ecological significance. Notwithstanding, ca and Southwestern Australia, there is we must mention the other papers concern- very little information that allows between- ing the relationships between herbivores continents comparisons. Perhaps one of the and plants in Chile. For example, there is reasons that defoliation of evergreen and experimental evidence that browsers can drought-deciduous shrubs has not attracted have an important role as killers of shrub much attention up to now, is that in general seedlings (Fuentes & Simonetti 1982, the amounts of leaf tissue lost to chewing Fuentes et al. 1983, 1986) and that grazers animals are small. In the Chilean matorral can be effective removers of herbaceous the average percentage lost is in the order plants (Fuentes & Le Boulenge 1977, of 8 to 10% (Fuentes & Etchégaray 1983), Jaksic & Fuentes 1980, Espinoza & Fuen- whereas comparable data for the fynbos tes 1983, Simonetti & Fuentes 1983). (South Africa) are only 2% (F. Kruger, These and other studies (Fuentes & Hajek personal communicatin 1980). Only in 276 FUENTES ET AL. Australia, Morrow (1983) found levels of 1977, Thrower & Bradbury 1977, Rundel defoliation in the order of 20-30%, but 1981 for descriptions of the Chilean these values seem to show high between- matorral). Defolating insects, the most years fluctuations (EF, personal observa- important leaf-eating animals, have their tion). peak abundances in spring (Atkins 1977) It is reasonable then to ask whether this and show only partial overlap with the defoliation has any biological significance growing season of shrubs (Fuentes et al. at all. At this point we should clarify that 1981 ). Leaf chewing insects are mostly our focus is only on the defoliation of large lepidopterans and coleopterans (Etchégaray (reproductive) shrubs and small trees. In & Fuentes 1980). There are no leaf-cutting our presentation we will: I) describe the ants in Chile. shrub defoliation phenomenon, 2) show Neither presently nor in the recent some of its patterns in space and time, and historical past there have been native 3) present the results of tests designed to browsing mammals that forage on large evaluate the significance of the phenomenon shrubs or trees (Simonetti & Fuentes in a community context. 1983). However, there is a list (Tamayo & Frassinetti 1980) of such large Pleis- tocenic herbivores, some of which might have had a role as defoliators of matorral 1HE PHENOMENON shrubs (Table 1). Notwithstanding their possible past importance, because of the Shrubs in the Chilean region with winter recent patterns of landscape modification rains and summer droughts exhibit their (Fuentes & Hajek 1979) and the relative vegetative growth from September through short turnover time of matorral ecosystems, October (spring) if they are drought- their visible effects nowadays are most deciduous, or from November through likely to be in some of the phenotypic January, that is from late spring to early attributes of the individual plants, rather summer, if they are evergreen plants than in their abundance patterns. One of (Montenegro et al. 1982). (See Mooney these attributes could be the presence of TABLE 1 Large herbivorous mammals in central Chile. Pleistocene (left column) and recent (right column) large mammals are indicated. The diversity in the past was higher than in historical times, but neither mammal abundances nor their importances as herbivores are known. Grandes marnfferos herbivoros en Chile central. Se indican los animales del pleistoceno (colwnna izquierda) y de tiem- pos historicos recientes (colwnna derecha). La diversidad fue mayor que en el pasado reciente, pero nose conocen ni las abundancias ni su importancia en cuanto herbivoros de estos animales. PLEISTOCENE HISTORICAL TIMES Order Edentata Megatherium medinae . • . • . • . • • • • . No analog Scelidodon chiliense • . • . • • . • . No analog Order Litopterna Macrauchenia sp. • • . • • . • . No analog Order Proboscidea Cuvieronius humboldti. • . • . • • . • . • • . • • . • . • . • . No analog Order Perissodactyla Equus curvidens. • • • • • • • • . • . • . No analog Equus sp. • . • . • . • . Equus caballus Onohippidium sp. • . • • . • . • • . • • . • . • • • • . No analog Order Artiodactyla Lama weddelli. • . • . • • . • • • . • • . • . • . • . • . • . Lama guanicoe Lama sp. • . • . • . • . • • . • • . • • . • • . • • . • • . • . • • • . • . • . No analog Antifer sp. • . • • • . • • • . •..••...••.. , •.•• , • • . Hippocamelus bisulcus No analog . , .......... , ..•• , , , ••.......... , , . Capra hircus CHILEAN TERRESTRIAL HERBIVORY 277 lignotubers in most matorral shrubs (Mon- PATTERNS tenegro et al. 1983). Although these sub- terranean burls have other explanations a) Species. Measurements taken in 1977, (Montenegro et al. 1983, Fuentes & Espi- 1978 and 1980 at Santa Laura on the noza 1986), it may well be possible that coastal ranges near Santiago (Thrower & defoliation of seedlings and saplings by Bradbury 1977), have shown that E(P) some of the large Pleistocenic herbivores is generally lower than 0.20 and that within could have been an additional selective a given year between-species differences factor for their evolution. Experimental are significant (Fuentes & Etchegaray 1983). defoliation of Quillaja saponaria seedlings Some species, such as Lithraea caustica or shows that they do recover even if this Quillaja saponaria, consistenly showed the treatment is repeatedly applied. Due to higher E(P) values, whereas shrub species lack of enough appropriate data regarding such as Baccharis spp., Muehlenbeckia other such attributes we will not discuss hastulata and Colliguaya odorifera always these issues further. exhibited E(P) values in the order of 0.01 The other currently important browsers to 0.02 (see also Fuentes et al. 1981 ). in the matorral are goats (Capra hircus), Other species, such as Kageneckia oblonga, which since their introduction have shown had values that varied in the range between to be capable of devastating some central 0.06 and 0.14. Chilean shrublands (Fuentes & Hajek b) Geographical variation. Within the 1978). Again, we will not discuss them area with a mediterranean-type climate here and interested readers are referred to there is species -as well as community- Fuentes & Etchegaray (1983). We will thus wide geographical variation in average E(P) concentrate all our attention on the relation values. For example, Colliguaya odorifera between defoliating insect and large shrubs. and Lithraea caustica show statistically Studies using tagged branches of shrubs significant positive correlations (P's < 0.05) have shown that: a) insects forage only on in their E(P) values with either altitude or newly produced leaves and frequently eat latitude of the collecting site (Fuentes & only a fraction of them (Fuentes et al. Etchegaray 1983). Kageneckia oblonga has 1981 ), b) the probability of insect-attacked been shown to exhibit an E(P) of 0.07 and non-attacked leaves of being dropped