1867.] Relations of Yerlasca. 171 



ed with pollen of the purplish violet, or normal form, the average 

 in this being 25 seeds per capsule ; then follows the unions with 

 pollen of the white variety, the average of seeds being in these 21 

 seeds per capsule ; and lastly in the unions with the variety with rose- 

 coloured flowers, the fertility of V. lychnitis, lutea, is reduced to 

 the low average of 18 seeds per capsule. Thus judging by the degrees 

 of fertility, we clearly see that the natural functional co-relations of 

 these plants in place of being regnlated by their respective colour 

 affinities, arrange themselves in an entirely independent and opposite 

 scale ; the extremes in the scale of colour given, viz., the purplish- 

 violet with yellow, manifesting the nearest functional co-relation. 

 -Again as a further complication we find that the white and yellow unions, 

 — the most closely allied of the colours mentioned, — hold a medial 

 position between the purplish violet and rose. How obviously futile 

 then, we may well remark, would our d priori conclusions have been, 

 as to the degrees of fertility of the above unions, on a presumed 

 coordination between colour and function in the phenomena of 

 hybridism ! 



It would thus appear from the results given in the foregoing tables 

 that in the hybridisation of varieties of distinct species characterised 

 by differences of colour alone, no definite relations whatever can be 

 observed between the affinities of colour, and the degrees of fertility, 

 but, that in these cases as in the reciprocal hybridisation of pure 

 species, the relative fecundity is a most variable and unpredictable 

 quantum. This view seems to me to be further supported by the 

 results of my experiments on the reciprocal hybridisation of the di- 

 morphic species of Primulas* in which I showed that the laws of dimor- 

 phism were limited in their action to the unions of the two forms of a 

 species ; the heteromorphic and homomorphic unions of distinct species 

 proving irregularly the more fertile. From considering the important 

 functional co-relations of the two forms of dimorphic species, and their 

 trifling morphological characteristics, together with the specifically 

 limited extent of their operations, we have less reason to be surprised, 

 if a similarly limited relationship should ultimately prove to regulate 

 the degree of fertility of those unions of differently coloured varieties 

 of a species as in Verbascum and analogous cases. Indeed, judging 

 * Linn. Soc. Jour. Vol. 8, p. 78. 



