150 
contact with the limits of the animal kingdom. See Agardh Classes, p. 4. and 
Martins, H. R. M. p.l76. For observations upon the nature of this irritabi- 
lity, see Dutrochet sur la Motilite, Paris, 1824, in which the author endea- 
vours to shew that the motion is the effect of galvanic agency ; and the same 
writer’s Nouvelles Recherches sur V Exosmose, in which he alters the ex- 
planation of the manner in which galvanism produces the motion, adhering, 
however, to his opinion of that subtle principle being the real agent. This 
ingenious naturalist might have been satisfied with attributing the phenome- 
non to an inherent vital action, without puzzling himself wfith a vain search 
after first causes, which always leaves the most successful inquirer exactly 
where he set out. For remarks upon the order in general, see De Candolle’s 
valuable Memoire, pubhshed at Paris in 1825-6, in one thick volume 4to. 
The relation that is borne by this order to Chrysobalanaceae will be explained 
under that order. To the tribes formerly included under the name of Tere- 
bintacese, Leguminosse are nearly allied in many important circumstances, but 
are distinguished by their stipules, which nevertheless exist in Canarium 
among Burseracese, and which do not exist in Sophora, a Leguminous genus. 
With Xanthoxylaceae they are allied through Ailanthus. The monadelphous 
stamens, iiTegular flowers, occasional simple ovary, style, and stigma of Poly- 
galacese, are all so many points of affinity with Leguminosae. 
In many respects this order is one of the most important which the bota- 
nist can study, but especially as it serves to show how fittle real importance 
ought to be attached to dehiscence of fruit in determining the hmits of natural 
orders. What may be called the normal fruit of Leguminosae is a legume, 
that is to say, a dry simple ovary, with a suture running along both its mar- 
gins, so that at maturity it separates through the middle of each suture into 
two valves ; but every conceivable degree of de\dation from this type occurs : 
the Ai'achis and many more are indehiscent ; Detarium is drupaceous ; in Car- 
michaelia the valves separate from the suture, which remains entire, hke the 
replum of Cruciferae ; in all lomentaceous genera, such as Omithopus, the 
valves are indehiscent in the line of the suture, but separate transversely ; in 
Entada a combination of the peculiarities of Cannichselia and lomentacese oc- 
cm's ; and, finally, in Hsematoxylon the valves adhere by the suture and split 
along the axis. The divisions which have been proposed in this extensive 
order are of unequal value ; it is possible that two of them, namely, Mimo- 
seee and Caesalpiniese may deserve, as Brown seems to think, the rank of 
sub-orders ; for they really appear to be of the same importance with refe- 
rence to Papilionacese, as Amyridaceae, Connaraceae, Anacardiaceae, and Bur- 
seraceae, with respect to each other. I give them, however, as I find them in 
De CandoUe. 
His first and most important division depends upon the form of the 
embryo, out of which arise the divisions called Curvembriae and Rectem- 
briae ; viz. — 
CURVEMBRI^. 
Radicle bent back upon the cotyledons. 
Tliese are distinguished into two tribes by the structure of their flowers, 
viz. — 
Tribe 1. Papilionace^. R. Br. 
Calyx with distinct kbes. Stamens perigynous. CoroUa papifiona- 
ceous. 
Tlie germination of this tribe varies thus : — some of the species push them 
cotyledons above ground, which become green, resembling leaves ; and of these 
none bear seeds which are eaten by man or animals : others germinate with 
their cotyledons under ground, and it is among these only that all the kinds 
which bear what we call pulse are found : the former De Candolle calls Pliyl- 
