8 INTRODUCTION. 
improbable that any one who makes a serious and prolonged study of the genus will 
fail to agree with the statement that this section is not real, but imaginary. The pods 
of species of the postulated section by no means agree with each other: their only com- 
mon feature is a negative oue; they are not samaroid pods, as is the case with the bulk 
of the Dalbergias. Their nearest allies are not Sissoae or Dalbergarieae or Triptolemeae 
in particular, but are to be found scattered throughout the genus. Some have pods 
with a thick corky mesocarp, others have little or none of this suberous thickening; 
still other species, eg., D. stipulacea, combine a local suberous thickening of the 
mesocarp opposite the seed with a samaroid structure elsewhere, Finally, апа 
perhaps most important consideration of all, the structure of the pod in all the 
specics of Selenolobium or Ecastaphyllum is such as to adapt the organ for dispersal 
by water or for the protection of the contained seed in swampy forests, and just in so 
far as it is possibly the result of environment does this structure fail to afford 
a character that is taxonomically useful. However, even after the species of 
Sélenolobium are rolegated to their natural positions within the genus, matters do 
not appear to be entirely satisfactory. Bentham’s section Triptolemea is a useful and 
a fairly natural section; his section Dalbergaria is even more natural and quite as 
useful, But Bentham’s Sissoa is residual rather than natural, since it includes 
many species that are precisely Dalbergarias except that they have monadelphous 
instead of isodiadelphous stamens; many species that are practically Triptolemeas except 
that they have larger flowers and usually longer styles; and finally, many species 
that, agreeing in the main with the monadelphous species which in other respects aro 
exactly like Dailbergaria, differ from these as regards vexillum as much as these differ 
from Daibergaria as regards stamens. If, then, the species of Dalbergia must be divided 
into sections at all, and when the crowd of species to be dealt with is considered, such 
a subdivision seems very desirable, these sections ought clearly to be groups of as 
nearly as possible equal natural rank, When such a subdivision is made, five sections 
must be recognised. These sections aggregate themselves naturally in two larger groups 
that may һә treated as subgenera, The characters on which this classification 
depends are detailed in the systematic part of this review; it is sufficient for the 
moment to say what they are and to explain what relationship they bear to the 
groups proposed by previous writers. All the Jcastaphylla of America and Africa 
appear to be Dalbergarieae as to flowers, One of the Asiatic Ecastapiylla (D. Вессатй), 
however, is certainly and the other (D. Albertisii) is probably а Triptolemea. Three of 
the Selenolobia are undoubtedly JDalbergaricae, and there is possibly a fourth belonging 
to the same section, Two of them are, however, Tripfolemeae, and in all probability 
a third belongs to the same section, The remainder belong to that section of the 
old and undifferentiated Sissoa, where the flowers are neither like the flowers of 
Triptolemea nor like the flowers of Dalbergaria, These species thus disposed of, the 
genus Dalbergia is easily divisible into two subgenera: (a) 8515504, of which the 
well-known species, D. Sissoo, exhibits in a well-marked fashion most of the character- 
istics; and (b) Amermnoy, of which D. Amerimnum may be taken as the type. This 
SissoA does not correspond exactly with the section of that name proposed by 
Bentham in 1851 or with the corresponding subgenus recognised by Daker and Kurz in 
1876, since it eliminates all those species that have monadelphous stamens, but have 
at the same time hastate wing-petals. AMERIMNON, on the other hand, does not (apart 
altogether from the inclusion of certain Selenolobia) correspond with the Dalbergaria 
