208 Campbell : The embryo-sac of PandAnus 



was available. I must here express my deep appreciation of the 

 great help afforded me by Professor Treub and the Director of the 

 Garden, Mr. H. J, Wigmann, who gave me every assistance in pro- 

 curing specimens. Pandanus offers special difficulties in securing 

 material, as most of the species do not flower freely, and even under 

 the extraordinarily favorable conditions at Buitenzorg only a small 

 number of species flowered during my three months' stay. 



Many of the specimens growing in the garden at Buitenzorg are 

 still unnamed, as the limits of the species are still very poorly un- 

 derstood, and the large genus Pafidamts seems to be very much in 

 need of a* thorough revision. Three of the species collected, how- 

 ever, were named, two of them being comparatively small species : 

 P. Artocarptis Griff., and P. affinis Kurz. The other was the large, 

 wide- spread species P. odoratissinnis L. f., which extends as far as 

 the Hawaiian Islands. The material was fixed in several ways : 

 one per cent chromic acid, weak Flemming's solution, and acetic 

 alcohol. All of the methods proved satisfactory. The younger 

 flowers offer no difficulties in the way of fixation, but in the later 

 stages the outer part of the carpels becomes hard and this must 

 be cut away so as to expose the softer tissues surrounding the 

 ovule. Owing to the small number of inflorescences that were 

 available, it was impossible to obtain any of the stages subsequent 

 to fertilization, but it is hoped that material may be procured which 

 will supplement the account of the earlier stages which is given 

 here. 



All of the Pandanaceae are dioecious. The pistillate flowers in 

 Pandamts are in dense heads, which in the smaller species are 

 borne several together at the end of a short branch, each head 

 being subtended by a conspicuous bract. In the larger species, 

 where the head of fruit maybe as big as a large pineapple, the heads 

 are solitary and enveloped in a large number of bracts. The Pan- 

 danaceae have been placed close to the Spar^aniaceae, and the 

 heads of flowers, as well as the spiky fruit look very much indeed 

 like those of Sparganium. In the smaller species ; e. g.^ P. Arto- 



r 



carpus (fig. i) the flower consists of a single carpel, which, in form 

 as well as in the position of the ovule, resembles that of Sparga- 

 fuiivi simplex^ but the conspicuous scale-like bracts that surround 

 the carpel in Sparganium are absent. In the large species like P. 



