98 Mr P. Sewell on the Morphology and [sess. li, 



two-celled. These cases suggest the origin of axile from 

 parietal placenta?, and though requiring further support from 

 other similar orders, they seem to emphasise more clearly the 

 affinities of Scrophulariacere with such orders as Gesneracea:', 

 where the two parietal placenta? yet remain constant. 



In regard to the stamens, we may remark that their 

 didynamous condition is directly the outcome of the bilabiate 

 form of tlie corolla, both of wdiich modifications secure the 

 better fertilisation of tlie flower. 



We may select genera, however, which retain the fifth 

 posterior stamen found in the allied order Solanaceae, as 

 Verhascum; others which show, as Scroi^hularia, its presence 

 at times as a fertile stamen, at times as a staminode, or it 

 may be entirely absent. 



Other genera show in addition the partial suppression of 

 the anterior stamens — Gratiola, perhaps, in as interesting a 

 manner as any — inasmuch as it serves as a link to the con- 

 dition in Veronica, where all trace of other than the lateral 

 ones lias disappeared. 



Miiller * has pointed out that where the stamens are 

 collected to the upper part of the flower, the missing stamen 

 is rudimentary, or altogether absent, apparently making room 

 for the style, which protrudes ahove the anthers. Where 

 the stamens are not so collected in the upper lip, as in 

 Scrophularia, an evident staminode, or even perfect stamen, 

 may be found; in such a case Miiller shows it will not be in 

 the way. He therefore considers that the disappearance of 

 the posterior stamen is to be accounted for by the natural 

 selection of tliose forms which are fertilised the most readily, 

 as above shown. 



It would at first sight appear that the abortion of one 

 stamen rather than another might be accounted for, not by 

 natural selection, Init by reason of the unequal growth in the 

 development of the parts of tlie flower. The results of this 

 inequality of growtli during development are very clearly 

 seen in Pinguicula, where, as Professor Dickson has pointed 

 out, the growth is from the anterior to the posterior, and 

 where the missing posterior stamen is tlius the last to be 

 developed ; wliilst, if developed at all, tlie two lateral appear 

 oidy as staniinodes. 



* Fertilisation of Floiccn, )i. 431. 



