Erythrococca and Micrococca . 601 
the scales may perhaps represent a modified androecium. That this sug¬ 
gestion is, at least sometimes, more satisfactory than the other, may be 
gathered from an examination of the arrangement which obtains in the 
plant on which the genus Poggeophyton , Pax (Engl. Bot. Jahrb., xix. 88), 
was based. This plant agrees with Erythrococca in having perulate buds, 
stipular thorns, and plumosely laciniate stigmas. It was believed, when the 
genus was proposed, that it differs from Erythrococca as regards its fruit. 
This, as we now know, is not the case ; the difference between Poggeophyton 
aculeatum and an Erythrococca , such as E. aculeata , is confined to the hypo- 
gynous ‘ disc That difference is certainly striking. In Erythrococca the 
‘ disc ’ is composed of small, flattened discrete scales, isomerous and alternate 
with the carpels. In Poggeophyton these hypogynous scales are replaced by 
large staminodes, isomerous and alternate with the carpels ; what Pax treats 
as the ‘ disc 5 is a ring of smaller oblong glands hirsute at their tips with 
long viscid hairs. The glands which constitute this ring are sometimes 
free, sometimes slightly connate below, and belong to a whorl external to 
that in which the staminodes are situated. The staminodes are obviously 
only somewhat imperfect stamens with stout cylindric-clavate filaments 
longer than the anthers; the anther cells are two in number, erect, free 
except at the base, and without pollen. The ‘glands \ on the other hand, 
are identical in appearance and character with the glands that form the 
extrastaminal ‘ ring ’ or the extrastaminal ‘ urceolum ’ in the male flowers 
of an Adenoclaoxylon , such as E. Kirkii or E. mitis , or of a Deflersia , such as 
E. abyssinica , E. acnleata , or E. subspicata. 
The difficulty presented by the conditions described is not, therefore, 
so formidable as it at first appears. The ‘ staminodes ’ of Poggeophyton 
aculeatum are identical in appearance as well as in position with the hypo¬ 
gynous scales of Claoxylon triste as described by Mueller (DC. Prodr., xv. 2, 
779), and with those of C. Menyharthii , Pax, except that in C. triste and 
C. Menyharthii these scales are merely cylindric-clavate filaments which are 
not surmounted by empty anther cells. Moreover, in C. Menyharthii , appa¬ 
rently as a casual abnormality, one or two small, linear-oblong, glabrous 
glands may occasionally be met with in addition to the staminodial hypo¬ 
gynous scales, though these glands are less conspicuous than the corre¬ 
sponding ones in Poggeophyton aculeatum , and do not, as in that species, 
form a complete hypogynous ring. 
Two other species, which also are confined to West Central Africa, 
agree in general facies with Poggeophyton aculeatum , and are evidently 
members of the same natural group ; in both of these species, however, the 
staminodes characteristic of P. aculeatum are replaced by ovate, subacute, 
flattened hypogynous scales of the normal Erythrococca type. One of the 
species, E. Laurentii , is readily distinguished from the others by having 
leaves of a different shape with shorter and stouter petioles, but the second, 
