41 



elusion that they originated from two different immigrant ancestors. The agents 

 responsible must again have been birds or perhaps wind currents. The seeds of 

 Lobelia are so minute that they could easily be taken up in the upper air cur- 

 rents and thus carried across great areas. Their stations as we have seen are 

 exposed, and swept by strong gales and as they are already at considerable 

 elevations their fine seeds could easilj- be carried by high winds which blow 

 almost continuously over these boggy summits with such force that the writer 

 found it most difficult to wallc. To the writer it seems much more feasible to 

 attribute the presence of Lobelia proper to wind currents rather than to bird 

 agency. The fruits of the true Lobelias being capsular they would not attract 

 birds to such an extent as baccate fruits would, thcnigh there is the possibility 

 of the seed having arrived in dried nuid on the feet of migratory birds. The 

 genus Ljobclia proper forms the crowning stage of the ijlant-stoeking of this 

 island group. Probably long after the present endemic genera, there arrived 

 species of tlie widely distributed genus Lobelia, which in the course of millen- 

 niums produced the endemic species of today. We see the same case repeated 

 in the Coinposifae. The oldest denizens of that family are undoubtedly the 

 Silversword Argyroxiphium, Hespi ruinainuu, and ^Vilkcsia, while Coreopsis 

 iCamptjlotheca) and LApochaeta are the youngest, and it is with these that the 

 genus Lobelia may be compared as far as length of residence is concerned. 



THE GENUS TEEMATOLOnELTA. 



In the year 1S91, A. Zablbruckner established the genus Trematoearpus on 

 what was then known as Lobelia maerosleielnjs Hooker et Arnott; the specimen 

 was collected Ijy Wawra on Kauai, which represents however a variety of the 

 species occurring on Oahu and jMolokai. Owing to the name of Trematoearpus 

 being pieoecupied Dr. Zablbruckner suggested the name Treniatolobelia which 

 was published by the writer in a College of Hawaii publication.* The genus 

 Trematolobelia differs from the true Lobelia in the capsule, which, instead of 

 dehiscing at the top into loculicidal valves, does not dehisce at all like a Ijobelia 

 but disseminates its seeds through oval or round holes in the capsular wall which 

 become larger towards the base of the capsule; the depressed umbonate vertex 

 remains however intact. 



The genus Trematolobelia possesses only one species T. inacrostachys which 

 occurs on Oahu, Molokai and Lanai, on the crests of high mountain ridges but 

 below the habitat of Lobelia Gauelichaudii. It is usually epiphytic but also ter- 

 restrial. It is a remarkably handsome species and besides ditfering from Lobelia 

 in the characters above mentioned, differs also in the horizontally branching 

 inflorescence. The flowers are slender and pinkish in color. On Kauai it is 

 represented by a very distinct variety (var. Kauaicnsis), occurring on the high 

 plateau in swampy forests, along stream beds, and at the summit bog Waialeale 

 in company with Lobelia Kauaensis, and its var. villosa, besides Drosera, Sani- 

 cula, Plantago, Compositae, Geranium, etc. In the mountains of Kohala on 

 Hawaii it is represented b.y another variety much more robust and with large 

 obovate leaves, larger white flowers and long racemes. This var. grandifolia 

 grows along the edge of Waipio, Alakahi and Kawainui gorges where there is an 

 enormous precipitation, at an elevation of about ■4,000 feet. 



* College of Hawaii Bulletin No. 2:4.5. plates 11 & 12. 1913. 



