DIALYSIS. 115 



flowers connected with the abnorma cucumbers already 

 cited. 



In a proliferated floret of the goat's-beard (Trago- 

 pogon pratense) Avhich had become virescent, the 

 usually syngenesious autliers were quite free. In 

 the normal floret of Compositge and the flower of 

 Cucurbitaceae the anthers are postgenitally, not con- 

 geni tally, united. 



In the abnormal axillary " stamens " of E^iphorbia 

 described by Schmitz there were frequently present so 

 many as four to six loculi, all of which he ascribed to 

 the anther of a single terminal axial stamen. But 

 Oelakovsky points out that in Euphorbia there are 

 really two stamens present, and an abnormality of 

 Schmitz's shows this, although that author placed, 

 as Oelakovsky shows, a quite wrong interpretation 

 thereon. In the axil of one of the leafy bracts was a 

 structure consisting of two practically opposite foliar 

 organs, one of which bore two interior pollen-sacs, and 

 clearly represented a stamen ; at the base of the other 

 leaf occurred two minute leaflets, which Schmitz 

 ascribed to the terminal bud of the whole axillary 

 structure, but which Oelakovsky says must represent 

 the two pollen-sacs of the second leaf. If this, as is 

 probably the case, is the right interpretation, then we 

 have a dialysis of the " single stamen " into its two 

 actual constituents, which are at the same time 

 virescent. 



In virescent flowers of Li/simacMa, Cyclamen, and 

 Anagallis it has been observed that the stamens 

 become completely free from the petals and inserted 

 on the receptacle ; in otherwise normal flowers of 

 Primula vulgaris and P. sinensis the same thing has 

 been seen (PI. XXXIX, fig. 9) ; this abnormality is of 

 great importance as showing that the petal and its 

 superimposed stamen are to be regarded as two 

 distinct and independent foliar organs fojp:nng two 

 whorls ; it is quite antagonistic to the view of 

 Duchartre and Pfeffer, founded on the altogether 



