THE MAKING AND UNMAKING OF FLOWERS. 12B 



limb, the calyx follows suit and becomes a tube as well, and supports that 

 of the corolla. 



Now, the bee alighting on the front petal, it is clear that its weight is 

 not equally borne by all parts of the tubes, " the strain will be felt " as 

 bearing heavily upon the anterior side, tending, as it were, to split the 

 tube across laterally. Let us see what nature has done to meet this 

 contingency. The calyx not infrequently has actually become two-lobed 

 in consequence, as in Salvias (fig. 51) and in the Furze, of another family 

 [LeguminoscB) ; and in order to strengthen the calyx strong ribs of woody 

 tissue are run up just where the strain is most felt, and where there is a 

 tendency to split the tube. Thus at the sides two new cords are added, 

 as well as in front ; and in some species, where the pull or drag would seem 

 to be especially great, an extra cord is inserted, as ^ 

 in the accompanying diagram of the cords in the 

 calyx of a species of Salvia. The calyx being 

 composed of five coherent sepals, each being a repre- 

 sentative of a leaf, will have a mid-rib. Let d stand 

 for these primitive dorsal cords, as they are prominent 

 at the back of the leaves, then five d's would be the 

 normal and complete number. But we find marginal ^ 

 cords inserted as well (m), only one on each of the posterior lines of union 

 where the strain is least, but two at the anterior and lateral lines of union, 

 s being a supernumerary cord in front. 



In other members of the Labiate family slightly different numbers of 

 cords are inserted ; but always such as are required in each case to meet 



m 



m 



m 

 m 



m 

 m 



m 



m 



iG. 54. — Corolla of regular Salvia 

 laid open, showing four perfect 

 ' stamens. 



the necessary strain caused by the habitual visitors of the flowers in 

 question. 



Next let us come to the stamens. The normal number is four, one 

 having been suppressed of the ancestral five. This number was undoubtedly 

 the original one, for the flower was once perfectly regular. This is proved, 

 not only by inductive evidence, but by actual cases where Labiates 

 "revert " to the original form, and bear regular flowers with four or even 

 five complete stamens (figs. 53, 54). Of these four, two arise from the 

 posterior and two from the anterior side of the corolla tube, but all four 

 lie along the back of the flower, as in the Deadnottle (fig. 50). The 

 anterior pair have become, presumably, pushed across by the head of the 



