Ecological zonation and the butterflies of the Sierra Nevada de Santa Marta, Colombia

Summary

	(1) A survey of the butterflies of the north Colombian mountain range, the Sierra Nevada de Santa Marta, was carried out by the author and Mr. George Bernard in 131 working days in 1971 and 1972, at 43 main sites and their environs from sea-level to 4000 m. The data used are the records of the 501 species and subspecies collected by the author.
	(2) The Sierra Nevada rises from low-lying plains, nowhere higher than 350 m, and the Caribbean Sea, to 5775 m, and it is separated from the Andean stem. It is only about 10 000 square kilometres in area, and with its well-defined borders it is an ideal study-area for a naturalist.
	(3) Restrictions imposed by the time of year spent in the range, by the impenetrability of some of the forest, and by the fast rate of travel, are discussed, as well as the advantages of this kind of semi-superficial survey.
	(4) The land from the outskirts up to 5000 m is divided into eight distinct zones, based on vegetational morphology, climate and altitude, and each is described in note form.
	(5) The analysis of the data is in two main sections. The first calculates the proportion of the 13 families' species endemic to each of the zones, and a ‘Coefficient of Relative Diversity’, by which between-family and between-zone comparisons may be made without reference to their total complements. From this the relative preferences of the butterfly families for particular types of vegetation and ranges of altitude are ascertained.
	(6) The second section introduces three coefficients of association between pairs of zones, tabulates the results using the butterfly data, and shows the cluster diagrams produced by them. The first two, the coefficient of Jaccard and the coefficient of maximum similarity, both exaggerate, in different directions, the effects of differences between the total species-complements of the compared zones, which, except in one case, represent differences in numbers of available niches rather than in thoroughness of recording. The third coefficient, that of Cole, is preferred, as it takes into account more data, and treats differences in total complements in a less exaggerated way.
	(7) In general, the associations are given low values by the coefficient of Cole, showing that the different conditions in the zones are accompanied by largely different butterfly faunas. This is likely to be due not only to restriction of the larval foodplants' distribution but also to sensitivity to certain conditions inherent in the butterflies themselves, and some instances of the former restriction appear from the analysis. Some of the values are discussed, zone by zone, and particular emphasis is laid on the origin of the open vegetation of the two zones between 250 and 1800 m in the south-eastern mountains of the range, in which the slopes are typically bare of forest. From evidence concerning the inhabitants, past and present, the climate and the topography, and from the analysis of the butterfly data, it is maintained that this river-forest pattern has been a permanent feature, and is not the result of recent deforestation by the inhabitants.