49 
here again we have a character which, taken by ino points : 
two different lines of descent, as Wrightia and Holar rhena have 
otherwise so little in common that Wrightia has been placed in 
Echitidece wae cnet in Plumeriee where it occupies a 
somewhat anomalous position. We should not fare better with 
any other otataeter if tie alone, although anyone might forma 
convenient basis for a purely artific ial arrangement. “Echitidew, 
like the majority of Twbiflore, are rich in ill- -defined genera, an 
in instances of parallelism which makes it so difficult to trace their 
phylogenetic Suioeeers To do this satisfactorily is beyond the 
scope of the presen paper, as it would involve a critical revision 
of the whole tribe ut We thitidece; for it is only from a thorough and 
Se tpl investigation into the str ucture of the genera com- 
posing the tribe that we may hope to solve the intricate problem 
of their “alee relationship. 
f I may venture to suggest a place tor Patt it would be 
near Wright ia. The suggestion is not new; be was erp by 
Blume in "eat pater iv. p. ne on account of general re mb ances, 
and by Mie his essay “ On the guclindios of "Son th: 
America,” p. ‘9, on wobowie of the great similarity of the fruits 
and seeds. I have already pointed out the homology of the basal 
tuft of the seeds vat Wrightia and of the basal awn of Kickria, 
and the practical identity of the structure of the 0 
both genera. Neither of vere characters by itself is of very 
great taxonomic importance; but when they appear combined, 
and coincide besides with a se ssninedl parallelism in the structure 
of the flower and fruit, they become indicative of a closer 
relationship of the genera. This is, indeed, to a certain degree 
the case with Wrightia and Kickxria. Certain Malayan species 
of Wrightia approach Kickxria rather closely in general 
appearance, and Blanco was actually mislead to enumerate Kickzria 
Blancoi as a species of Anasser,a synonym of Wrightia, Still 
there remain these differences:—the xstivation of the corolla 
lobes is in Wrightia the reverse ‘of that in Kickxia ; the corolla 
is divided down % the insertion of the stamens and (with, I 
believe, a single exception) provided there with variously anaeed 
appendages, instead of surrounding the staminal cone with a cup 
or bell- shaped widening of the tube; and, finally, t there is Ae 
ie 
surrou r th eu , 
described lately two new genera from Cochin-China, Microchonea 
and Paravallaris, which belong possibly to the same stock as the 
resi species of Wrightia and Kickwxia, so far as I can judge 
owering specimens. As to Funtumia, however, the re- 
edit hlancs with Kickria ends with the homology of t mi 
appendage and the practical identity of the structure of the 
embryo. I have, so far, so in vain for another genus in 
Mo. Bot.Garden 
1906 
‘ 
