22 Plant Invasion on Lava Flows. 



the dense koa forests where Acacia koa is the prevailing tree, with 

 a thick undergrowth of ferns, labiates and innumerable other spe- 

 cies, are limited with remarkably few exceptions, as far as the 

 forests of Kona are concerned, to the pahoehoe. 



From any hill the traveler through this district can pick out the 

 aa flows from a distance by the lines of ohia (Metrosideros poly- 

 morpha) forests, which divide the koa into sections. When he 

 arrives at the rocky flow he will find a totally different character of 

 undergrowth, where such plants as Polypodium pcllucidum (folded 

 form), Lycopodium con mini, Vaccinium penduliflorum , Raillardia 

 and certain other plants are conspicuous. From this, I believe 

 that the koa forest is the final type for this region, and when one 

 finds an aa flow supporting koa, it must be relatively very much 

 older than other aa flows in the vicinity. 



In later years, however, outside influences, especially cattle 

 grazing, have considerably changed the floral aspect of the coun- 

 try in certain places. On account of greater ease for penetration, 

 a more succulent undergrowth, and perhaps more available water, 

 cattle have worked their way into the forests on the pahoehoe and 

 destroyed the undergrowth. In a few years the introduced flora 

 becomes established, and a park-like meadow, usually in every 

 sense an ideal pasture, is the result. However, the indigenous 

 trees, which after many years have become adapted to the dense 

 undergrowth, epiphytic creepers and moist soil, are much weak- 

 ened, and are not able to withstand the new, more xerophytic 

 conditions. The weakened trees eventually fall an easy prey to 

 destructive insects, but as far as I was able to observe were not 

 touched by the cattle. 



The cattle penetrate to a much less degree on the aa owing to 

 its rough character, while the more xerophytic undergrowth is 

 less succulent. When they do, the character of the flora is less 

 changed for the reason that complete destruction of the under- 

 growth is impossible, while this ohia, which has not become 

 adapted to such wet conditions as the koa, does not succumb to 

 insect attack. Ohia varies tremendously with the habitat, but 

 any great change in conditions would probably affect it the same 

 as koa. 



From the preceding paragraphs it must not be inferred that 

 ohia forests are not found on some pahoehoe flows; for the damp- 



