241 
The greater part of the specimens were collected at Banda on 
the slope of the Goenoeng Api. It is a conspicuous faet that such 
a great number of specimens have regenerated from fragments. 
Partially this is undoubtedly due to the stones which drop con- 
stantly from the mountain into the sea. In one very restricted 
locality the diver collected 36 regenerating fragments, most of 
which were of small size. A couple of these are represented in figs. 
93 — 96 . The parts of the mother-corallum with nearly parallel 
septa and costae are easily to be distinguished in the centre of 
the regenerated specimens, whilst the new formed parts have septa 
and costae which are arranged radially round the centre of the 
original fragment and therefore these new septa and costae are in 
some places directed at right angles with those of the central part. 
In further growth of the colony these new parts enlarge and the 
colony usually assumes an arched form. Then the only differences 
with normal colonies are found in the central part, where a cen¬ 
tral calicle is wanting and some of the septa and costae have an 
abnormal course. A number of such regenerated specimens of H. 
philippinensis are described by van der Horst as the new species 
H. louwinae. I have compared these coralla with a number of 
specimens of H. philippinensis in my collection and they prove to 
be quite identical. Some of van der Horsts specimens were 
also collected in the Banda Islands and their costal spines have 
a few grains at the top (cf. van der Horst 1921), a characteri- 
stic which is invariably found in all specimens of H. philippinensis 
in my material. The mode of growth of the colonies of the Siboga 
Expedition is the typical one found in regenerating specimens. 
In one regenerated specimen with a longer diameter of 342 
mm a large bud has developed at the lower surface as an irregu- 
larity of growth where two regenerated parts of the colony have 
fused (fig. 127). It is attached with a broad base to the colony, 
the larger diameter of its marginal part is 44 mm. Its calicle has 
divided into two parts by the fusion of two opposite septa. This 
young individual may be a true lateral bud, but it is also possible 
that it has developed in a similar way as the false buds which 
are described in the following pages in Halomitra robusta. 
True lateral buds are found in one very regularly developed 
specimen with an obvious central calicle. At the lower surface of 
Vidensk. Medd. fra Dansk naturh Foren. Bd. 79. 16 
