143 



ahuudant material of fruiting' specimens in all stages of development iu the 

 Wawra collection, permitted me to study most thoroughly the structure and the 

 development of the capsule. I shall describe the actual facts more thoroughly, 

 first because it necessitated the geuerical separation of Lobelia iiiacrostacliij>i 

 Hook, et Arn. and second liecause this mode of dehiscence of this plant is the 

 only one known witliin the family CaiiipaiiuJaceac as understood by Bentham & 

 Hooker. 



"Immediately after anthesis the corolla drops oft" and the staminal tul)e with 

 the basal part of the enclosed style breaks someAvhat above the conical insertion. 

 The calyx remains however on the receptable. The capsule assmnes gradually a 

 woody consistence through a thick zone of sclerenchym fibres which form lieneath 

 the parenchymes of the receptacular wall. It appears that Hillebrand consid- 

 ered this stage as the mature capsule. When the development of the scleren- 

 chymal zone has reached a certain point, then the epidermis of the receptacular 

 wall becomes separated from the woody part of the latter and with the epidermis 

 also the parenchymal layers beneath it. At the same time occurs the dropping 

 oft' of the calyciue teeth. During the process of separation there are formed in 

 the lateral woody part of the capsular wall, between the strongly protruding 

 ten ribs, oval or circular holes, which find their origin in the even accumula- 

 tions of sclerenchym fillers. The number of holes is variable, and becomes 

 greater at the base of 1lie capsule, also the presence of the same is more con- 

 stant at that place, between each two ribs. In all other Lobrlioidcae the recep- 

 tacular wall remains adnate or united with the carpids even at maturity of the 

 fruit; different, however, is the ca.se in Trematocarpus. Within each ci41 of the 

 capsule, the carpids separate from the woody capsular wall by means of shrivel- 

 ling and finally become torn. The small seeds, which in the meantime became 

 fully mature, reach the vacant space between the capsular wall ;ind the carpids, 

 and are dispersed from the capsules through the movement of the plant ; dispersal 

 is facilitated through the downward bent position of the capsule. 



■•Occasionally there appear during the drying of the capsule, on the e(iually 

 ligneous vertex, fine slits or furrows; these very rarely extend the entire thick- 

 ness of the vertex wall, and are even then nuich too small to permit the passing 

 of seed. These slits or furrows evidently caused Hillebrand to assume that the 

 capsule does occasionally open with fine slits at the apex ( 'indehiscent or opening 

 by small pores at the vertex"). 



"The seed is small, about 1 nnn long, usually ovoid, or weak pyramidal, red- 

 dish brown and smooth; laterally the seed is surrounded l)y a lighter-colored 

 duplicature of the testa which is almost square in shape ; this margin is formed 

 through the miitual pressure of the numerous ovules. The testa is composed of 

 a single layer of cells; these are elongate to polygonal, are arranged in longi- 

 tudinal rows, and their walls are minutely perforated. The albumen consists of 

 relatively large polygonal cells, in cross section arranged in four to five rows; 

 their contents are rich in fat. The embryo is straight, the radicle thick spindle- 

 shaped, constricted below the semicircidar lobes of the seed. The pollen is glo- 

 bose and exactly as in Lobelia." 



For the criticism and replies by Bolting IIemsle\, Stapf and ZahlViruckner, 

 see Annales of Botany, Vol. VI:1.')4 (1892), Vol. vfl :289, 396 (1893). 



