4 E. B. Williamson 



region of marked seasonal changes there are more or less frequent changes 

 in the dragonfly fauna of various habitats due to unusual or abnormal sea- 

 sons. It is obvious that the premature drying up of a pond might result in 

 the death of the larvae of several species. An unusually high water stage 

 of a river might submerge all the resting places for emerging larvae at the 

 time when they are due to emerge, or might coat larvae with a deposit of 

 stifT clay, resulting in the wholesale destruction of many species. Late frosts 

 in early summer or a premature freeze in the fall might cause the death of 

 imagoes and thus at least temporarily eliminate certain species from certain 

 habitats. Unusual seasons may also operate to the advantage of certain 

 cxtralimital species. For example, an early and continuously warm spring 

 may permit the extension northward of a southern species, and it is probably 

 this factor, associated possibly also with favorable air currents, that some- 

 times carries the southern Panfala hymenaea as far north as Ontario. 



In the low-lying American tropics temperature ceases to be an important 

 factor in itself in determining dragonfly distribution. Drought takes its place. 

 Drought makes itself felt even in the well-watered and forested equatorial 

 belt, though, due to permanent rivers and lakes, its effect is not universal 

 over a region as low temperature is. It is obvious, with low temperature 

 absent as a limiting factor in the seasonal distribution of dragonflies, at least 

 as imagoes, that more numerous adjustments in other factors are possible in 

 the life histories of these insects. For example, in the north the rigors of 

 winter can be bridged by a dragonfly only in the egg or larval stage, but in 

 the tropics it is not impossible that the period of drought might be passed as 

 an Qgg, a larva, or an imago. The existence in the tropics of both temporary 

 and permanent bodies of water have resulted in many diverse adjustments 

 in the life histories of dragonflies, both as larvae and imagoes, to these periods 

 of drought. For example, warmth and abundant food obviously might permit 

 the shortening of larval life, while other factors might permit the indefinite 

 extension of imaginal life. When, as is usually the case, temporary and 

 permanent bodies of water occur in close proximity in a region, the result 

 is a rich and varied dragonfly fauna. 



It is interesting to note that in the tropics many of the most successful 

 and specialized dragonflies are those which depend on, or are able to endure, 

 a temporary water supply (wet weather ponds or streams) for their larval 

 life, and that in the north our highly speciahzed species are pond and lake 

 (not stream) dwellers. 



Too little is known of the detailed life histories of neotropical dragon- 

 flies to allow us to cite particular examples of the various adjustments indi- 

 cated as possible in the preceding paragraphs. This is true of the species of 

 Erythemis, and yet enough is known to permit of a few positive statements 

 and some warranted conclusions. The larvae of the species of Erythemis, 

 so far as known, are pond or lake dwellers, always so, possibly, in the case 

 of some species, and frequently or usually in the case of all. The adults 

 have a long seasonal range, occurring on the wing apparently throughout the 

 year, and are apparently as numerous, if not more numerous, during the dry 

 season. Since the larvae of some, if not all, may live in temporary as well 



