the proportions are as 23 to 29, relatively to the pure unions of the latter. Thus, in whatever way we proceed, the general results are the same, testifying to the highly remarkable fact announced by Gärtner, that varieties of a species, characterised by no other differences than that of colour, are occasionally so differentiated functionally, that the cross-unions, as compared with the fertility of the pure unions, invariably indicate a certain degree of sterilisation! In connection with this higher relative fertility of homochromatic to that of heterochromatic unions, as limited to the crossing of varieties of a single species, I will venture to add that this law not only holds, but, as I believe, extends to and regulates the functional relations in accordance with the relative colour affinities of the varieties crossed. Thus for the sake of illustration, we may take the three primary colours of the cyanic series, namely, blue, violet, and red. Now beginning with red, we know that greater physiological changes must take place in the minute anatomy of the petals of an originally red-coloured flower to give the impression of blue than that of violet. Hence we might suspect that a species presenting varieties characterised by such differences in colour, would likewise afford different degrees of fertility in their conjunctive functional relations, the blue and red yielding less fertile unions, than the violet and red; while the violet holding an intermediate colour position between these, might be equally as fertile in its unions with the blue as the red variety. In practical illustration of these relations, we may take the results of the various unions of V. phaniceum and varieties given in Table 1. Thus the V. phoeniceum with purplish-violet flowers yields more seeds when fertilised by the pollen of the rose-coloured variety, than by that of the white variety, in the proportion of 5 to 4. Again the white variety of V. phœniceum fertilised by the pollen of the rose variety yields an average of 29 seeds per capsule, and by that of the purplish violet variety the average per capsule is 26, that is as 9 to 8, in favour of the unions of the rose and white varieties. We see here evident co-relations between the degrees of fertility and the colour affinities of these plants in their respective sexual unions, and I venture to look for more marked differences in these respects, had we as subjects of experiment, varieties of a species presenting three, or at least two, of the primary colours with intermediate shades irrespective of the white. The latter being rather unsatisfactory from its similar relations to the primary colours, though in such instances as the above of the purplish-violet, rose and white, in which we have secondary colours forming intermediate steps between the primary and white, by a gradual dilution of the colouring principle, we find that the white, agreeably to the above views, form less fertile conjunctions with the violet than the rose-coloured flowers. Before passing from this point of my subject, I will now only add that I have thought these indications of a tangible law, co-relating and regulating the sexual functions of varieties. when crossed-dim though they as yet undoubtedly are- worth noticing, as we are as yet in utter ignorance of anything like definite or specific laws in these phenomena, the results being considered as most capricious and uncertain. Gärtner's second proposition is, that in the hybridism of differently coloured varieties of distinct species of Verbasca, the conjunctions of the similarly coloured flowers are more fertile than these of dissimilarly coloured flowers. For example Gärtner shows that on the calculation of V. lychnitis, fl. alba, yielding with its own pollen 1.000 seeds, it yields when fertilised with the pollen of V. blattaria fl. alba, 0.622 seeds, and with that of V. blattaria, f. lutea, only 0.438, so that the similarly coloured unions of these species are more fertile than the dissimilarly coloured unions in the proportion of 3 to 2. Let us now see then in how far this law of the differences in the fertility of the homochromatic relatively to the heterochromatic unions, is borne out in the case of my own experiments as given in the preceding Tables. And here again, for the sake of clearness, and facility of reference, I will restate them in a tabular form, and show as clearly as possible the differences in the relative fertility of the homochromatic and the heterochromatic unions, in each case, by making calculations from an assumed 100 seeds produced by the more fertile of the two unions compared. The results may be conveniently arranged under three heads; thus, first, the heterochromatic unions, or those in which the unions of differently coloured flowers are the more fertile: second, the homochromatic unions, or those in which similarly colour * Versuche über die Bastarderzeugunj, 1849, section 216. ed flowers are the more fertile and lastly, the irregular unions in which no relations are observed between the degree of fertility and affinity of colours. A. 1.-HETEROCHROMATIC UNION, the MORE fertile. 2 3. 4. 5. 6. 7. 8. 1. V. lychnitis, lutea, by pollen of V. blattaria, alba, 100 to 91 100 to $3 1. 2. 3. 4. B. 2.-HOмOCHROMATIC UNIONS, the MORE fertile. V. lychnitis, lutea, by pollen of V. thapsus, lutea, 6. C. 3.-DEGREE OF FERTILITY AND AFFINITY OF COLOUR IRREGULAR. 1. V. lychnitis, lutea, by pollen of V. phœniceum, 2. V. lychnitis, lutea, by pollen of V. phoeniceum, 3. V. lychnitis, lutea, by pollen of V. phoeniceum, In A. and B. of the above comparative tables, I have ar ranged those unions in which a certain regularity is observed between the colour relationship and the degree of fertility. Now, by comparing the 14 unions therein given, we find that the heterochro matic unions are, in the greater number of cases, more fertile, viz., as 8 to 6, than the homochromatic unions, and that this higher fertility, in every case, results from those unions in which the yellow variety of the species is treated as female. Again that the average proportion of the four heterochromatic to the four converse homochromatic unions in the first of the above tables is nearly as 7 to 6 in favour of the former. In B. 2 of the tabulated results, we see in one instance the homochromatic unions with yellow as female exceed in fertility the converse heterochromatic union; but in the other cases given in lines 3 and 5, this higher fertility of the homochromatic unions is yielded by the white variety; the relative proportions of these being much more marked than in the above cases of the heterochromatic union with the yellow variety as female, viz., as 4 to 3, whereas, as we have seen, in the heterochromatic, A. 1, the proportions are as 7 to 6. In further illustrations of this point we see in B. 2 that the yellow homochromatic union of V. lychnitis, lutea, by pollen of V. thapsus, lutea, relatively to the heterochromatic unions of the former with pollen of V. thapsus, alba, is nearly as 5 to 4, so that we here again see (as in the heterochromatic and homochromatic unions in A 1) a more intimate approximation between the products of these two unions, than occurs in the other cognate unions of B. 2, in which the white variety is the more fertile. These curious relations, however, as I have already shown, are partly explained by the fact, though we can only dimly see why it should be so, that in the pure unions of the white and yellow varieties of the above mentioned species, the white, in every case, yields more seed than the yellow; whereas in the cross-unions the yellow variety in general is the more productive. But, it may be asked, how is the greater potency of the pollen of the white variety relatively to that of the yellow variety, as shown in the above tables to be accounted for? Does it really imply that the female element of the yellow variety yet retains its normal or original potency, the male element alone having become absolutely less potent, as compared with the male element of the white variety. This hypothesis, analogically considered, does not seem to me at all improbable. I think we have clearly seen by the comparative results of the pure and mixed unions of the yellow variety with those of the white, that the pure unions of the yellow do not yield a degree of fertility at all proportionate to that of the like unions of the white variety, as judged by the relative fertility of their cross-unions; and that accordingly this would seem to be due to an acquired weakness in the generative powers of the yellow variety. In noticing this point in a former part of my paper, I treated it as if both sexual elements had undergone a similar decrease in their generative powers; but we here see that it is more particularly, if not altogether confined to the male element. Now, as the results of hybridisation show that the pollen is more susceptible to the concomitant sterilising action of hybridism than the female element, may we not suppose that the debilitating effect of continued self-impregnation will also manifest itself more quickly in the male than in the female element, and thus afford an explanation of the decreased sexual powers of the male, as compared with the female element, in the yellow varieties of the above species of Verbasca furthermore, showing us that as it has been a slowly acquired quality, so will it be in its elimination and regainment of its pristine vigour. The relations of the several reciprocal unions in the above tables is another point which we must briefly consider, as having most important bearings on the subject of our present enquiry. A hasty examination suffices to show that these are much complicated. Thus V. lychinitis, lutea, in its two unions with the white and yellow varieties of V. blattaria, the heterochromatic unions are the more fertile; whereas in its two unions with the white and yellow varieties of V. thapsus, we find it yields the more fertile by a homochromatic union. Again V. blattaria, lutea, in its four distinct unions with the white and yellow varieties of V. thapsus and V. lychnitis, yields the higher degree of fertility in the heterochromatic unions, while the V. blattaria in its similar unions with the white and yellow varieties of V. thapsus and lychnitis is, singularly enough, more highly fertile in the homochromatic than the heterochromatic unions. Lastly the V. thapsus, lutea, yields more seed by its heterochromatic unions with pollen of the V. lychnitis, alba, than by its homochromatic unions with the V. lychnitis, lutea; whereas in the converse unions we have seen that the V. lychnitis, lutea, is more fertile in the homochromatic unions with V. thapsus, lutea, than in the heterochromatic unions with V. thapsus, alba! The tabulated experiments given in C. 3, afford another source of complexity to the question under examination, inasmuch as they are quite irregular in the relative degree of fertility produced by the affinity of colour. Thus by the three unions of V. lychnitis, lutea, with pollen of the three varieties of V. phoeniceum, the most highly fertile is that in which V. lychnitis, lutea, is treat |