1907] HOLM—RUELLIA AND DIANTHERA 313 
ends obtuse or acute; their surface, however, is always very distinctly 
granular. Viewed in longitudinal sections the cells containing cysto- 
liths are usually shorter than the sclerotic. The cortical parenchyma 
of Ruellia thus shows a very characteristic structure.’ 
Inside the endodermis is a thin-walled, continuous pericambium, 
which surrounds the mestome with its mechanical support of stereome. 
There are four strands of stereome, each consisting of five to twelve 
typical stereome cells in one or two layers between the pericambium 
and the leptome. The leptome itself shows the presence of raphid- 
ines, but they seem to be very scarce in this species of Ruellia, 
since I never found more than one in each cell, and only in a very 
few places. Raphidines were first described by Russow,® who 
observed them in the root, stem, and leaf of Hexacentris coccinea, 
etc. VESQUE® has also described them as characteristic of Cyrtan- 
thera, Meninia, Adhatoda, and Fittonia; while Router (/. c.) found 
them in a number of species of Thunbergia. They are also to be 
observed in Mendoncia and Pseudocalyx, according to RADLKOFER. 
The hadrome contains numerous vessels which extend to the 
center of the root, but secondary formations had already commenced 
so that the primitive organization of the stele could not be ascertained; 
however, the four strands of stereome may indicate a tetrarchic 
structure. Although the specimens of Ruellia were collected as late 
as the third week of September, no deposits of starch were observed 
in the roots. 
Besides the thick, somewhat fleshy secondary roots, there are 
some that are quite thin on account of the less development of cortex, 
which consists only of three strata, thin-walled throughout, with no 
7 A cortex of exactly the same structure has been described and figured by HENRY 
G. GREENISH (Pharm. Jour. and Trans. London 1891:839) as representing that of 
the root of what he took to be Phlox carolina. However, the root of that species does 
that I have examined. The rhizomes and roots which GREENISH had before him, 
d which 1d th £ Sp Cem Mt i 
a plant that has about the same - distribution as Spigelia, and whose rhizomes some- 
what resemble those of Spigelia. Hence the statement that cystoliths occur in the 
root of Phlox pa Se (SOLEREDER, p. 622) may be safely omitted. 
8 Sitzungsber. Naturf.-Gesells. Dorpat 5:308. 1878-1880. 
9 Sur quelques formations cellulosiennes locales. Ann. Sci. Nat. Bot. VI. 11: 
I. 
