170 PALMACE, 
an elevated central line, to which the costz reach, and down 
which they are decurrent a short way. 
Pinnules tough, coriaceous, upper and lower ones much re- 
duced and narrow, upper not awned, but with an obtuse emar- 
ginate coriaceous point. 
Capitul. femin. axim inflorescentia terminans, bracteis 
peduncul. crassa, ex axillis mascul. solitar. exserent valde соп- 
cave, interdum bi-tripartit, axim vel peduncul. nempe om- 
nino cingent, et latississime insertion, insertae margine super- 
currente. 
Bractee immediate capitule stipantes, basi late, margin. 
supercurrenti, extime bilobæ, interior integre  ovato-lan- 
ceolate vel sublanceolatee acuminate. Pauce angustiores. 
Pistilla antepenultima suffultient. 
Pistilla dense aggregata cuneato-obovata, pressione varie an- 
gulata, parte 829 vel dimid. super. supra angulata, pyrami- 
dalia, ad apicem szpius faceium inferiori alequando superiori 
rima glutinose hiante (Stigmatosa.) 
The irregularity in situation or direction of the stigmata at 
once suggests that the carpella are not solitary, and although 
it would be improper to say positively, that they are not, 
merely from adult specimens, yet I believe they may always 
be recognised as ternarily arranged. 
One thing is obvious, that the diversity in the situation of 
the stigmatic chinks, points out that the same carpella are not 
always developed. In Pandanus too, the stigmata all have the 
same direction which is another point of difference. 
The appearance of the stigmata and the pistilla, is not very 
ünlike certain instances of true Gymnospermous ovules. 
All the bracts are deflexed from the base, and are thence 
quickly ascending. 
This view is confirmed by the fact, that if a set be selected iu 
which the stigmata point (as in all such tases they do) to- 
wards each other, no perianthial scales will be found between 
them. 
1- — l-óvulata, ovulo erecto-anatropo tegumento 
