Plant Invasion on Lava Flo7vs. 21 



2. Lower cryptogams become established on the aa at an 

 early date, and eventually cover the flow to a considerable extent, 

 some of these species being rather rare on the pahoehoe. 



3. Ferns and phanerogams onh^ become established on the 

 aa a long period of time after these become established on pahoe- 

 hoe of the same age, other conditions being the same. 



4. The plants to be found on the new flow are the same as 

 those found on older flows in the immediate vicinity. Poly podium 

 pellucidii})i , Sadlcria cyathcoidcs and Metrosideros polymorpha were 

 usuall}' the prevailing plants at all points and elevations studied 

 on the pahoehoe, while a species of white lichen was usually com- 

 mon on the aa. 



5. A fertile soil is apparentl}' formed in the cracks of the 

 pahoehoe sooner than amongst the aa particles. 



Mctfosidcros polymorpha is one of the important plants which 

 prepare the way for the establishment of manj^ plants on the flow. 

 The roots spread over the smooth portions of the flow, often from 

 one crack to another, forming pockets to catch a soil formed of 

 dead leaves and other debris. 



In viewing the older flows it is rather difficult to ascertain 

 their relative ages, for the reason that the Hawaiian flora assumes 

 its mature form at an early date. In passing over many hundreds 

 of flows the prevailing characteristics of invasion were found to 

 be somewdiat as shown in the following paragraphs. 



In the scrub region above the wet forests there is no great 

 distinction between the flora of the pahoehoe and the aa. The 

 latter, however, has a greater proportion of lichens, occasionally 

 these plants occupying the aa to the exclusion of all other vegeta- 

 tion. The prevailing plants are Cyathodes Tamciamciae , Dodonaea 

 viscosa, scrub Metrosideros polymorpha, Sophora chrysophylla and 

 other characteristic mountain plants. Large portions of either 

 variety of flows are bare rock, and when a soil is formed it is usually 

 a ver>' thin layer, limited almost exclusively to the pahoehoe. 



In the dense forests of the rain}- belt there is usually a good 

 soil on the pahoehoe varying from six inches to a few feet in 

 depth, sometimes only the tops of the billows remaining in sight 

 to record the character of the flow. The aa, however, with few 

 exceptions, still retains its rough file-like character. I find that 



