PORTULACES, AND THEIR ALLIES. 67 
posed as a distinct genus urider the name of Dufourea, might, on 
account of the capsule, be better referred to Cerastium, where 
Fenzl once placed it. 
Holostewm is a small genus, most elaborately described by J. 
Gay, and reducible, as he proposes, to two or even to a single 
species. It has the capsule of Cerastiwm, but may be maintained 
as distinct on account of the habit and inflorescence and the pecu- 
liar seed. This is flattened from front to back as in Dianthus 
and its allies, but the radicle, instead of being short and straight 
as in those genera, is turned down in a projection of the inner 
face by which the seed is attached, thus combining the two forms 
of embryo which prevail in Caryophyllee. 
Stellaria is a large and widely spread genus, tolerably natural, 
and, as to the large majority of species, well marked by the three 
styles, bifid petals, and the capsule divided to about the middle 
into as many entire or bifid valves. But there are a few anoma- 
lous species, mostly isolated or nearly so, whieh have been sepa- 
rated into distinct genera upon real or fancied discrepancies, which 
however we think ought, from the general concordance of charac- 
ters, to be retained in Stellaria. These are— 
l. Larbrea, A. de St. Hil., founded on S. uliginosa, which has 
the petals and stamens more distinctly perigynous than in most 
other species, though still very slightly so ; but this is a question of 
degree only, as a more or less distinct perigyny may be observed 
in several other species where the petals are much reduced. 
2. Malachiwm, Fries, has been generally adopted for the 5. 
aquatica, placed by Linneus in Cerastium as having 5 styles. It 
differs, however, from that genus in the styles.being alternate with, 
not opposite to, the sepals. The habit, petals, &c., are those of Stel- 
laria nemorum ; the capsule only differs in the valves being rather 
less deeply bifid ; and the number of styles is, in Indian specimens, 
not unfrequently reduced to three as in other Stellarias. 
3. Krascheninikowia, Turcz., was adopted by Fenzl as distin- 
guished by the petals emarginate only or shortly bifid, although 
the original Siberian K. rupestris is apparently identical with the 
Carpathian Stellaria bulbifera, and very nearly allied to some 
other eastern or South- European species. The genus has, how- 
ever, since been remodelled by Maximowiez and made to rest on 
dimorphous flowers, the apparently perfect ones in the East- 
Asiatie specimens being usually sterile, whilst the seeds are pro- 
duced by small, almost apetalous oligandrous flowers near the base 
of the stem. But this, although, as far as I am aware, the first 
r2 
