52 PEOCEEDIXOS OP THE 



connection between it and the Angiospennous gynoocium, so we 

 postnlatc'd a liypotlicticiil group, tlie lleiiHangiosperuis, as the 

 ilirt'ct ani't'stors, chaiactcrised l)y having open instead of closed 

 carpels with llie ovules marginally borne. We introduced the 

 prefixes jxv) and fit to distinguish respectively tlie anthostrobilus 

 with open cari)els and that with closed carpels. The Angiosperins, 

 then, as a wliole, possess eu-anthostrobili and the Hemiangiosperius 

 l)ro-anthostrobili. 



In a second joint paper (6) reasons were brought forward for 

 regariling the individual "flowers" of the Giietales as very 

 reduced pro-anthoslrobili, and we theorised to the effect that the 

 existing members of this puzzling group might be held to be 

 verv aberrant survivors of tiie lI(>miangios[)erms. 



It is important to distinguish clearly between the two sides of 

 the Strobilus Tlieorv, viz. : what may be termed respectively the 

 ]}ennetlitalean and Kanalian sides. The Ilanalian hypothesis can 

 stand alone, even if the supposed J5ennettitalean relationship 

 were disproved. Some botanists apparently have failed to grasj) 

 this. To clinch my point 1 had myself discarded Eiigler's theory 

 and embraced a Eanalian origin of^\ngiosperms some time before 

 I was aware of the existence of the Bennettitean fructification. 



Possibly some botanists have been converted to the idea of a 

 Kanalian derivation of Angiosperms by seeing in this view a 

 |)lausibl9 origin of Flowering Plants from ancestors akin to the 

 Pennettitales. Kealising the possibilities of the strobilus theory 

 of the Hower, thev are not likely to return to the idea of the 

 primitiveness of one consisting only of a single stamen or carpel, 

 even if the supposed Bennettitalean relationship may have to be 

 relinquished. 



Tlie BeimettltaJean side. The crux of the strobilus theory as 

 regards its Bennettitalean side centres round the antliostrobilus, 

 tluit is to say, the sequence in which the sporophylls are borne on 

 the axis of the cone. Invariably both in Angiosperms and 

 Bennettitales the megasporophylls are borne on the axis above the 

 microsporophylls. AVitli the exception of the Gnetales {]]ehrit- 

 schm)—o. case of especial significance— no other group of plants, 

 fossil or recent, is known characterised by such a strobilus. All 

 other Gymnosperms when cones are present have these unisexual. 

 In heferosnorom forms of the Lycopnda, however, bisexual cones 

 apiiear to be the rule; but here we find that the relative position 

 of the two kinds of sporophylls on the axis is the reverse of that 

 in the anthostrobilus— the basal part of ihe cone being female 

 and the upper male, e. .g, CalamostacJnjs, Lepidostrohus,SclaiitneUu. 

 Prom the foregoing it becomes manifest that the type of cone 

 possessed by the Bennettitales, and named by us the anthostro- 

 bilus, is at present unique among fossil plants. On the assumption 

 that the Bennettitales have descended from Pteridospermous stock 

 the question arises :-Has the anthostrobilus arisen from this 

 i)lexus at more points than one? It may have, but m the 

 existing state of our knowledge it is permissible to imagine that 



