88 



mediates of which have become extinct, k'aviiig- one or two species of that par- 

 ticular type on some of the ishuids. It is quite possible and probably true 

 that a i;(i()(l many cither species of that type existed, whicli, if we had any 

 kuowledue of their characters would link them together with those of other 

 sections. 



SECTIONS DEI.IStSEOTDEAE. JTIUTELLAE, AND PILOSAE. 



The first section was established by llilli'l)i'and and is characterized by the 

 minute calycine teeth. It may be argued, and it is true, that the calycine lobes 

 of the Hawaiian Lvbduiidcnc are the most variable character, though in this case, 

 especiall3' in this section, while the calycine teetli are minute, the corollas are 

 so much alike, of the same shape and color, a.s well as being arranged on more or 

 less long drooping racemes, that it would be unnatural not to group them 

 together. 



The first section uinnbers now six species, against ten in Hillebrand's ar- 

 rangement ; three have l)een classed with section hirtellae, and one very little 

 known in Hillebrand's time, with the section pilosae. The three former belong 

 to a section which the writer was justified in establishing by the discovery of a 

 number of new species, (four in number, to which were added two by other 

 authors) all of which occur on the island of Kauai. They are quite distinct in 

 habit of growth and in corolla, whose gener;d character runs through all nine 

 species. 



The species of the section hirlclldr. have (piite a different habit of growth 

 from the species of the other sections, are usually single-stcmuied plants and 

 only in one or two instances occasionally branching from the base. This brings 

 the species of this section closer to the section paUiicK formes, and the writer would 

 look upon them as miniature forms of the lattei-. The leafy crowns, which are 

 of course terminal, are much smaller than in those of section pahnae formes. 



It may also lie remarked that the species of section delisscoidcac occur in the 

 outskirts of the forests on the leeward sidr and at hiwer altitudes, l()()()-3()00 

 feet, not in the swainpy forests, while those of the section liirtcllac occur in the 

 wet forests from 3000 feet up to otHln feet, in the humid rainforests, save a few 

 eicceptions, occurring in lower altitudes, but still in the rain forests on the wind- 

 ward side. 



Anyone knowing them from field experience cannot help but arrange them 

 in a section by themselves ; this would perhaps not be done by a person only 

 acquainted with them from herbarium specimens. Of course it would sinq)lify 

 matters very much indei'd would one adopt the method set forth by II. Baillon in 

 his Histoirr dcs phnils where he throws all niir Hawaiian endemic lobelioideous 

 genera togethei- iiitu our genus {Delissra), dividing the latter into as many sec- 

 tions as there existed genera. That is one w;iy out of the diiificulty but to the 

 writer's mind, does not lead to a better undo'sanding of this striking element 

 of our Flora. Tlir (h'eper we go into the study of these plants, dividing them 

 according to tlieii' natural affinities, tin- moi-e we will learn in regard to their 

 ancestry, evolution. ag<'. etc., more than by simply throwing them all together 

 into one genus, without furthei- thought. These plants must be studied in the 

 field: only then can one come to iirojiei' coiiclnsions. 



