• 38 STKUCTUPvE AND AFFINITIES OF CALLlTRICHACE^. 
coat, having the ordinaiy appeanince of the extenial tunic where two 
(primine and secundine) are present, through which the nucleus, which 
. is opaque, may be very distinctly seen. The formuen is also distinctly 
marked, having a slightly fringed or indented margin, and the canal 
leading from it to the pointed apex of the nucleus is in most instances 
visible. It is of an oval form, the foramen not being terminal but 
bent inwards towards the placenta, bringing to mind Caryojjhyllacea, 
and from the point of its attachment, wliich is near the middle, de- 
scends tbe short raphe. Prom this time up to the maturity of the 
seed this tunic is always distinctly visible, and, as before noticed, forms 
in the seed a small wing, extending from the hilum to the micropyle. 
The radicle does not point directly to the micropyle, the latter being 
halfway between it and the hilinn. 
The position of CaUitrkhe appears to me to be between CanjopJiyl- 
tacem and ElatinefB on the one hand, and Batideee on the other (as the 
ovary of Bails consists of 2 carpels, as in CaUitriche right and left the 
axis, with 4 cells), and the latter being so near Ferbenacea that it may 
almost be regarded as a section of that family (Trans. Linn. Soe. vol. 
xxii. p. 412), will account for the similarity of the ovary and ovule of 
CallUricJie to that of the Borarjine^. In the stamen posterior this 
genus agrees with the Pododcmcnecs, where in Caslelnaria the two 
stamens are posterior ; Cadelnavia has also numerous glands on its 
flattened stems closely resembling thof'e of tins genus. 
In the opinion of those botanists who would place CaUitriche near 
Ilaloratjacer, I am unable to concur, for how can a genus having the single 
stamen posterior (in Hippuris \i is anterior*) an ovary of preei'^ely the 
same structure as Borar;;„er,, and having the same appearance, except 
that the walls are thiimer, be well compared with Ilaloracjaceas? The 
ovule and seed also more nearly reseral)le those of the Boraginere. (es- 
pecially where the cells are nearly erect and closely approximated) than 
those of any other family I recollect to have examined, except perhaps 
Mollugxne,2, among which Ademgramma has an ovule attached by its 
side with a superior foramen. In both these the raphe, as in CallU 
tnche, IS next the placenta, but in lUppuri, and other Ilaloragace^ as 
far as is known, it is lateral, and in the nearly-allied Onagrace^, when 
* In Onagracece the sinsile stamen, if I remember rnrTP^fW u i.t, 
m:cc of b.in, posterior, but the fikm^nt is STy t wi ?e7 itff rold tl!T?tT; 
a fimgular phenoxiieaon. ^ ^"^^lea nau rouna the stjle. 
