400 DR. J. C. COSTERUS ON MALFORMATIONS 
gynecium are normal in double Fuchsias, it very rarely occurs 
that some of the stamens are affected by petalody." 
The above, however, does not preclude the presence of addi- 
tional stamens. We possess, for instance, a flower with thirteen 
stamens, consequently five more than the ordinary number. But 
instead of additional petals, we find here a great many thread-like 
appendages, as just described. Even the remarkable fact presents 
itself here, that the normal petals have not appeared at all. It will 
soon become manifest that by the latter circumstance, Fuchsia be- 
trays a tendency that may be regarded as proper to the whole genus. 
As to the relation between the thread-like parts and the other 
parts of the flower, it often happens that a thread-like element 
has grown together with a stamen, so that the latter part appears 
as shown by Pl. LVII. fig. 3. In this case the little thread may 
have a petaloid extension on the right side and on the left (fig. 1c), 
and in this form join the stamen. The sepals are also sometimes 
adherent for some little distance to the parts named, and in one 
instance the style even showed a thread-like appendage grown 
together with its lower part. In the latter case, however, the 
excrescence probably was a production of the pistil itself. 
The strangest additional parts occurring in Fuchsia are men- 
tioned by Masters; they are flower-buds on pedicels alternate 
with the petals. This phenomenon in flowers which are other- 
wise normal affords an instance of axillary prolification, whereas 
the additional petals, threads, &e. exemplify what is by Masters 
styled polyphylly. 
The cause of polyphylly is explained by Goebel in his paper on 
double flowers cited above. As for the Onagrariacez examined, 
he pronounces chorisis to be the cause. 
According to Goebel the primordial tubercles of a petal will 
divide into various pieces, the latter generally developing into 
separate parts. An incomplete division causes lobed petals, a form 
which we, however, found both in ordinary and in additional 
petals. Goebel considers a production of intercalated parts, quite 
independently of existing ones, very improbable, because a rami- 
fication ofthe tubercles can easily be detected*. This conclusion, 
however, does not tally with the fact that in double flowers the 
number of vascular bundles is greater than in single ones. This 
difference in number, though in many cases trifling, may be con- 
siderable in other flowers. In examining the calyx-tube of & 
single flower, we find only the vascular bundles ascending fo 
* Goebel, in Pringsh. Jahrb. für Wiss. Bot. 1886, p. 247. 
