86 CONTRIBUTIONS FROM THE NATIONAL HERBARIUM. 
its texture of the genus Dalbergia, and a flower which does not give any conclusive 
evidence except that the wings are apparently free. In the absence of better mates 
rial, especially of the fruit, the status of this species remains doubtful. The type 
was collected at Chilla, District of Puebla, Mexico, by Andrieux (no. 440). 
LONCHOCARPUS PHASEOLIFOLIUS Benth. Journ. Proc. Linn. Soc. Bot. 4: Suppl. 93. 
1860. 
Another species of doubtful status, known only from a description of incomplete 
specimens at Kew. Collected also by Andrieux, near Tehuantepec, Mexico (no. 462). 
LONCHOCARPUS PARVIFLORUS Benth. Journ. Proc. Linn. Soc. Bot. 4: Suppl. 89. 1860. 
Certainly a good species, according to the description, but unfortunately not found 
in the material at my disposal. In the key it should be placed near L. atropurpureus. 
LONCHOCARPUS LITTORALIS T. S. Brandeg. Zoe 5: 157. 1903. 
According to Brandegee, the only ‘‘generic variation is in the calyx, which, instead 
of being truncate, with very short obsolete teeth, is divided half its length into lan- 
ceolate lobes.’? More important, and, indeed, decisive, as diverging characters, are 
the dehiscent pods and stipellate leaflets. This plant does not belong in Loncho- 
carpus. 
: TRANSFERRED SPECIES. 
Gliricidia meistophylla (Donn. Smith) Pittier. 
Lonchocarpus meistophyllus Donn. Smith, Bot. Gaz. 56: 55. 19138. 
This differs from Lonchocarpus in the appearance of the leaves and flowers. More- 
over, the staminal tube is split, with the vexillar stamen often almost free, and the 
basal fenestrellze are either absent or with very thin margins; the calyx is quite 
aberrant, and the thickened margins of the ovary seem to indicate a dehiscent legume. 
For these reasons the transfer, as above, is proposed. 
Pringle 5649 and 6748, Rose 5869 and 9923, and Purpus 1187, labeled as Loncho- 
carpus, all have dehiscent pods and stipellate leaflets; they should be placed among 
the Galegeae, perhaps in or near Galactia. 
SOUTH AMERICAN AND WEST INDIAN SPECIES. 
Many Middle American specimens have been identified as Loncho- 
carpus violaceus Benth., but this name is now found to be untenable. 
The type of the genus, Z. punctatus H. B. K., was confused with the 
supposed violaceus by Bentham. To exhibit the distinctive characters 
of the two and to show how they differ from the Middle American 
species of the section Punctati, they have been included in the above 
key and are fully described and discussed hereafter. 
Among the species collected in Santa Marta, Colombia, by Herbert 
H. Smith and in Venezuela by J. R. Johnston and the writer some 
were found to be new, and others to belong to little-known types. 
Their detailed descriptions are also added to the present paper, as 
well as that of another species, discovered in Bolivia by H. H. Rusby 
and belonging to the section Fasciculati Benth. 
Lonchocarpus benthamianus Pittier. Figure 37. 
Lonchocarpus violaceus Benth. Journ. Proc. Linn. Soc. Bot. 4: Suppl. 95. 1860, 
not H. B. K. 1823. 
A small tree, the branchlets, leaves, and rachis of the inflorescences entirely gla- 
brous. 
