162 PEINCIPLES OP PLANT-TERATOLOGY. 



TEE THEORY OF ANTHEB-8TBUGTUBE. 



Morphological Structures of the Upper Surface. 



In some of the structural stages, as in those of 

 Dictamnus figured by Celakovsky, it is clear that two 

 of the loculi are represented by the leaf-margins, and 

 the other two by laminar outgrowths from the upper 

 surface of the leaf (PI. XLIII, fig. 7) ; in more advanced 

 stages of the metamorphosis these laminar outgrowths 

 dwindle in size and eventually disappear. He also has 

 figured, and the present writer has also observed, as 

 described in Vol. T of this work, similar outgrowths 

 from the upper surface of foliage-leaves in Phlox 

 paniculata (decussata.) (PI. XLIII, figs. 8-10), and they 

 are also found in Saxifraga ligulata (PI. XLI. figs. 5-7) 

 and the lower surface of the leaves of the Indian knot- 

 weed {Pohjgomim cuspidatum). 



Celakovsky describes a most interesting abnormal 

 leaf of Hieracium glanduloso-dentatum (PI. XLIII, figs. 

 11 and 12) which was forked at the apex for a short 

 distance, the inner margins of the two free lobes so- 

 formed being inrolled,and these inroUed margins were 

 continuous with those of two median ventral outgrowths 

 extending from the base of the two lobes to more than 

 half-way down the leaf ; this structure he explains as 

 due to dichotomy of the leaf having occurred as far as 

 the extent of the ventral outgrowths and inroUing of 

 the inner margins of the two lobes so-formed, and to a 

 subsequent fusion of the lobes as far as the present base 

 of the fork. At the base of this same Hieracium-leai 

 is a small ventral lobe, with its upper surface directed 

 inwards, which is quite comparable in position to the 

 ventral fertile spike of the Ophioglossum-Biporo'phjW. 

 Now he imagines that anther-structure might have 

 been brought about by just such a transformation in 

 the upper half of a leaf as has here been described, 

 coupled with a further dichotomy of the whole leaf so 



