250 Mr. J. W. Howell on the Structure of the Capsule 



generated upon the centre of the face of the carpellary valves, 

 is opposed to all that is certainly known upon the subject. 

 Even in those cases in which the ovules are indeed borne on 

 a great portion of the inner face of the carpel, the centre is 

 entirely destitute of them, as in Limnocharis. The first sup- 

 position, therefore, based on this untenable assumption, cannot 

 be entertained ; in fact, it is only mentioned to show the greater 

 necessity for receiving the remaining mode of explanation, 

 notwithstanding its apparent paradoxical nature, viz. that the 

 seminiferous dissepiments are true parietal placentas, but that 

 each stigmatic ray is double,— formed of the adjacent lateral 

 portions of the stigmas of two contiguous carpels ; the two por- 

 tions of the stigma of each carpel in the more complex capsules 

 of the higher species being separated by an intervening mem- 

 brane ! Ex. Argemone, Papaver, 



It is obvious that nothing less than the most rigorous de- 

 monstration will suffice to establish so remarkable a mode of 

 explanation, and this can only be effected by tracing the suc- 

 cessive steps of gradually increasing elaboration of the cap- 

 sule, from its most simple condition in Bocconia, through Mac- 

 leaya, Chelidonium, Glaucium, Hunnemannia, Eschcholtzia, 

 Meconopsis and Argemone, to the state of greatest complexity 

 in Papaver somniferum. Notwithstanding the great difference 

 in the forms and appearances of the capsules of these genera, 

 they exhibit a perfect similarity in all essential particulars of 

 their structure, their differences being dependent, not on the 

 relative disposition of their constituent parts, but on their 

 proportion, magnitude, and number. 



The simplest state of the capsule in the Papaveracea is ex- 

 hibited by Bocconia, Linn., in which it consists of two dor- 

 sally-compressed carpels united by their margins, forming a 

 flattened one-celled capsule containing a single seed, which is 

 attached to the inferior part of the replum or annular recep- 

 tacle formed by the united margins of the carpels, from which 

 the greater portion of the latter separates in the form of valves. 

 This annular receptacle is shown to be identical with true pa- 

 rietal placentas, although, except at a single point at its base, 

 it does not bear ovules, by the latter being developed through- 

 out its entire vertical extent on both sides the capsule in the 

 cognate species, Macleaya cor data, Brown [Bocconia cor data, 

 Linn.). The capsule is crowned by a deeply bifid stigma, 

 whose internally plumose halves being widely reflexed corre- 

 spond in situation and direction to their subjacent valves, and 

 therefore alternate with the intervalvular parietal placentas. 



It is interesting to remark, that in this, the simplest state of 

 the structure of the capsule, the relation of parts exemplifies 

 the law which expresses the necessary alternation of stigmas 



