On “ Fasciation .” 
61 
pomes are perfectly free, in another case they are congenitally 
fused, yet only partially so, in the third case this fusion has 
proceeded much further, so that the ovary of the smaller pome has 
been partially absorbed into the tissues of the larger one. I 
regard the fusion as due to the fact that reversion to the primitive 
corymbose fructification being only partly successful, and two fruits 
only having been produced, these, owing to either one or both of 
them approximating in size and vigour to the condition of the 
ordinary cultivated apple, have necessarily become congenitally 
united owing to the lack of space for the complete individualisation 
of each pome. The double pome represented in Fig. 1 came from 
a tree which never bore any other form of fruit. 
Post genital, mechanical or “ real ” fusion of parts is of com- 
paritive infrequent occurrence and is of little importance for my 
purpose ; for examples I may cite the syngenesious anthers of the 
Compositae, and shoots of Ivy, Elder, Beech, &c. 
II.—Positive D£doublement or Branching. 
This phenomenon consists in a congenital increase or multi¬ 
plication of parts or organs, and is especially well seen in many 
facts of floral morphology. The older writers were inclined to 
regard this as the main factor concerned in the evolution of the 
flower. But Celakovsky points out that it really only accounts for 
a comparitively small number of phenomena, the chief tendency 
being towards reduction in the number of parts. It is, perhaps, 
seen best in the ordinary “ doubling” of garden flowers ; as in the 
Rose and Daffodil. (I may here consider both normal and 
abnormal cases-together.) Positive dedoublement is almost always 
abnormal. Leaving out of consideration the acyclic types of the 
Ranales, I may, perhaps, regard those cases, as in Lythrum and 
Paris, where the members of each perianth-whorl are greater in 
number than the 5 and the 3 which are the respective typical 
numbers for Dicotyledons and Monocotyledons, as examples of 
normal positive doubling. A similar case which at once suggests 
itself is that of the formation of the pappus in Compositae (con¬ 
sisting as it does of great numbers of hair-like segments) from the 
primitive five-leaved calyx of the ancestors of that order. In 
abnormal flowers of various Monocotyledons, such as Crocus, Iris, 
Tulip, Snowdrop, &c., it commonly occurs that, either in alternate 
whorls or throughout the flower, four members are produced in the 
whorl instead of three ; and this as the observation of intermediate 
