1980] 
Young — Ecology of Cicadas 
193 
1972; 1980a, b; 198 la, c) although precise data on the abundances of 
skins in legume plots versus non-legume plots is still lacking. At 
Cuesta Angel cicada nymphal skin patches too were found beneath 
legume trees. If legume trees provide some form of highly suitable 
environment for developing cicadas in tropical forests, emerging 
populations each year will be spatially disjunct according to the 
spatial distribution of the legume trees whose root crowns provide a 
primary resource for developing nymphs. The suitability of 
Leguminosae for developing cicadas may involve both physical and 
chemical properties of the classes of root sizes exploited by various 
age-classes of nymphs. The observed low densities of nymphal skins 
in all of the plots at Cuesta Angel, relative to previously obtained 
densities of the same or similar species in other regions (e.g., Young 
1980a, b; 1981a), may therefore be a function of the very dispersed 
condition of the legume trees at this locality. A striking contrast is 
made with the association of nymphal skins of species such as Z. 
smaragdina and F. sericans in relatively large patches of adult 
Pentaclethra macroloba in nearby premontane tropical wet forest 
(Young 1980b). Densities of these cicadas range from 5.4 to 9.3m 2 in 
patches of two or more P. macroloba, estimates considerably 
greater than for the same species at Cuesta Angel. I interpret such 
observations to be the result of P. macroloba occurring as clumps of 
several trees thereby increasing the size of a single resource patch for 
cicadas, which results in either greater oviposition in the patch or 
greater survival of nymphs, or both. The river-edge plots in the 
Cuesta Angel study illustrate quite well such an effect. Such plots, 
although quite large, only contained one or two widely scattered 
legume trees and not clumps of such trees, and some did not contain 
legumes at all but were situated near such trees. The observed very 
low densities of cicada nymphal skins in these large segments of 
forest is due to an absence or scarcity of suitable root crowns for 
cicadas. The tree plots, on the other hand, although very small, are 
highly suitable for cicadas and therefore densities are high. 
The pattern of cicada nymphal skins being associated with legume 
trees in tropical forests can have other explanations as well, ones not 
involving a presumed coevolved interaction of the sort suggested 
above. For example, selective logging of tropical forests may leave 
behind the relatively soft-wood legumes thereby increasing their 
relative abundance as a resource for insects such as cicadas. Thus 
the likelihood for an ovipositing cicada to discover a legume tree 
