or TUE SEPALS IN SEUAPIAS. 493 



pentagonal, or tlie like. Here, since wc find that both in normal 

 and abnormal variation the fruits tend to assume similar types, 

 it is hard to resist the conclusion that the production of these 

 forms is due to tendencies which make a part of the constitution of 

 the plant ; for the disease which, in the case of tlie abnormal va- 

 riation, might be taken for the cause of the modification, is ab- 

 sent in the case of the normal variation. 



It w^ould appear, from Dr. Masters's * Vegetable Teratology,' 

 that petalody of the sepals is of rare occurrence, and that in- 

 stances of the assumption of the characters of the inner by tlie 

 outer perianth-segments of Orchidacese have not hitherto been 

 recorded. It would seem more intelligible if the two upper 

 petals had taken on the labelliform condition, as the labellum 

 is itself a modified petal ; and it is, at first sight, hard to form 

 even a conjecture wliy the lateral sepals of Serapias should so 

 systematically undergo this change. 



If, however, we turn to Mr. Darwin^s ^Fertilization of Or- 

 chids ' (p. 294 <fcc.), we learn something of the homologies of the 

 labellum, wbich may perhaps supply us with a clue. Mr. 

 Darwin, from a study of the course of the spiral vessels through 

 the tissues of the flowers of Orchids, arrived at the conclusion 

 that the two missing stamens of the outer whorl are combined 

 with the labellum on either side of its central nerve, the evi- 

 dence of which is found in the bundles of spiral vessels which 

 I hav^e called the lateral nerves. Now it is exactly these parts 

 that are brought into contact with the modified portions of the 

 lateral sepals ; and it does not seem a very improbable conjecture 

 that tbe change observed in these latter may be due to the 

 partial diversion or bifurcation of the spiral vessels belonging to 

 the two missing stamens. It is very tempting to make one 

 more suggestion, founded upon the preceding hypothesis, and to 

 speculate whether (if it be true that the introduction of spiral 

 vessels belonging to either stamen into the two lateral sepals 

 suffices for the production in either of half a labellum, in place 

 of the ordinary sepal-structxixe) the combination of the two 

 stamens with the anterior petal may not readily account for its 

 change into a labellum. 



I have twice seen forms of OpJirys insectifera^ L., in each of 

 which one of the flowers was entirely destitute of a labeilum, 

 and the lateral sepals were united along their anterior margins, 

 so that they assumed the position of the missing labellum. In 



