Structure. 85 



Apocarpiae : 

 Idioblasts rare in the filament, always present in the anther. Stomata absent from all 

 the surface of the filament (Brachyceras). Idioblasts absent from all parts 



of the mesophyll of the ovary and carpellary styles TV. caerulea. 



Idioblasts absent from filament and anther (Anecphya) N. gigantea. 



The dry microspores (pollen grains) of N. alba are oval and deeply 

 folded ion one side (Kerner, 1895, fig. 218); in Brachyceras (Fig. 38, b-e) 

 they are spheroid, with greatly flattened poles, or, in N. capensis, lentic- 

 ular, and the edges may be folded under from three sides. Fischer (1890) 

 speaks of them as folded in (eingefaltet) in N. dentata, lotus, rubra, and 

 alba, the infolded half being separated from the other by an equatorial line. 

 Along this line the exine separates in germination. The exine on the 

 infolded half is thinner and smoother than on the other half, and has 

 parallel to the equatorial line a ring of similar thickness and sculpture to 

 the outer half (Fischer, 1. c). In water the grains become spherical in a 



A 



Fio. 88. Pollen: a, young grain of S. flavo-virtns ; b e, N. caerulea; b, just 

 matured ; c, nearly dry ; d, e, as shed from the anther, seen from opposite sides. 



few minutes, and along the equatorial line in N. capensis, caerulea, zanzi- 

 bariensis, and hybrids the two halves of the exine separate slightly ; they 

 seem to be too small to cover the whole swollen grain. In pollen of N. 

 odorata mounted in glycerine jelly, some grains have the exine caps still 

 touching at one edge, but wide apart at the other, while some are equidis- 

 tant all round, and some are unbroken. Fine rods, points, and granules 

 cover the exine in Eu-castalia, and one-half of it in N. tetragona and flava 

 (Caspary, 1865 ; 1888). It is smooth in Brachyceras, and, according to 

 Caspary (11. a), in Anecphya, Lotos, and Hydrocallis. Fischer (1890), 

 however, found it granulate (kornig-walzig) in N. rubra and dentata, and 

 smooth in N. lotus, and I have found it smooth in N. mexicana. In 

 germination Caspary frequently refers to an operculum, and in one place 

 (1888) to the "smaller half" of the exine ; but the two parts which I have 

 noticed in Brachyceras and N. odorata are practically equal, and Fischer 

 speaks simply of two pieces (zwei Stilcke) in N. alba, and elsewhere of 

 the two " halves " of the exine. The term " operculum " is therefore 

 misleading. 



The pistil is composed of 5 to 35 carpels, the number being extremely 

 variable for any one species. They are placed radially round a central 



