
          Euphorbiaceae, so that either you have made some mistake or 
Michaux's two plants are one and the same. At [?] your 
plant has never a 3-celled capsule but a lenticularly compressed 
1-celled one, and the ovary although it has traces of 3 cells, 
yet two of them are abortive and without ovules.

4. Dolichos from Georgia. I have nothing like this in my herbm. [herbarium]. At 
first it appeared to me a Canavalia, but then the upper [tissue?] consists 
of two large rounded lobes. I therefore think your plant a true 
Dolichos ([with?] the fruit [opposo?], and [crossed out: what] you have sent does 
not not permit me to say whether it is [?] or compressed)
The upper lip is minutely 2-toothed, [crossed out: show] at the tip, showing 
that it is composed of two segments adhering at their base and 
nearly [?]. It is perhaps identical with some of the 
West Indian species of Swartz, but as to what really is a species 
this genus and what a variety, nobody seems to know. When I 
examined De Candolle's herbm. [herbarium] no. N825 after he had described the Leguminoseae 
not one species of Dolichos was named and in the Prodrs. [Prodromus] you will 
meet with the v. v. of v. s. sp. If proper characters were discovered 
I do not see that this genus would be worse than its allies, but 
at present no one can get one, on account of the very meager descriptions 
given previously. It was the part of the Prodr. fl. Pen. 
Ind. or. [Prodromus Florae Peninsulae Indiae Orientalis] at which I was most alarmed as I tried to get Hooker & others 
to do that genus, but no one would. Graham had undertaken the whole 
order of Leguminosae, and when at length he found his doctor's slips 
prevented him, I agreed to let him off if he would do the Phaseoleae 
or [?] Dolichos alone, but he laughed in my face at the mere 
idea of him or anyone else doing justice to that genus from dried 
specimens, or the [?] no. of species we have cultivated. What I did 
is before the world. It cost me more than a fortnight's [added: discretions any] cogitation, 
and I fear is not over satisfactory, although it enables anyone 
to understand the Peninsular species who attend to my [minute?] 
characters of the subjection. The specific characters must be thought 
in all cases almost entirely in the flowers & fruits.

And now as to my Torreya. I would have much rather 
preferred it having been published in your own country, and although 
you had drawn up the [added: short?] character in my [name?] and left 
your description to another occasion. All Friday & Saturday (this 
is Monday) I spent in confirming your analyses, which I do ad [?] 
I have altered some of your sketches, and added others, but then if you say

[cross writing]
I might presume but I say no! no! as matters stand.

Now there is a long letter for you and I now 
beg to conclude. With yours truly
G. Walker Arnott

Did I send you no. 1056 of Wight's Catalogue which 
I [have? cleared? a blunder?] form of Ammannia verticillata. I must 
have been dreaming, for it is identical with Amm.  [purpurea?]

No. 1026 I forget [tho?] whole his [?] of the plant, or how 
I [?] to make the blunder. Wallich's Am. Hyneana 
is a mere variety according to a specimen in Herb. Hooker.

Have you seen Spach's analysis of Lechea and 
some allied genera in Hooker's "Companion," one of the last nos.? 
He makes too many genera but his analyses are accurate. 
There are at least two genera [added: or subgenera] shown confounded [unidentifies?] 
differing very widely in the structure of the placenta but his 
supposed new species of Lechea: in one I have got from you 
& Greene and Boott a Lechea [in?] [?]. Among his rubbish 
there are more gems than Hooker or Lindley incline to admit. 
And at all [accounts?] he deserves credit for the accuracy of his 
analysis. This till now scarcely one botanist has accurately 
described the family of Lechea.
        