I am quite willing to suppose, with Professor Weldon (p. 248), that the pea _Stratagem_ may, as he suggests, be such a case. I am even willing to accept provisionally as probable that when two gametes, themselves of mosaic or blended character, meet together in fertilisation, they are more likely to produce gametes of mosaic or blended character than of simply discontinuous character. Among Messrs Sutton’s Primulas there are at least two striking cases of “flaked” or “bizarre” unions of bright colours and white which reproduce themselves by seed with fair constancy, though Mendelian purity in respect of these colours is elsewhere common in the varieties (I suspect mosaics of “false hybridism” among allelomorphs in some of these cases). Similarly Galton has shown that though children having one light-eyed and one dark-eyed parent generally have eyes either light or dark, the comparatively rare medium eye-coloured persons when they mate together frequently produce children with medium eye-colour. In this connection it may be worth while to allude to a point of some practical consequence. We know that when pure dominant--say yellow--is crossed with pure recessive--say green--the dominance of yellow is seen; and we have every reason to believe this rule generally (_not_ universally) true for pure varieties of peas. But we notice that in the case of a form like the pea, depending on human selection for its existence, it might be possible in a few years for the races with pure seed characters to be practically supplanted by the “mosaicized” races like the _Telephone_ group, if the market found in these latter some specially serviceable quality. In the maincrop peas I suspect this very process is taking place[163]. After such a revolution it might be possible for a future experimenter to conclude that _Pisum sativum_ was by nature a “mosaicized” species in these respects, though the mosaic character may have arisen once in a seed or two as an exceptional phenomenon. When the same reasoning is extended to wild forms depending on other agencies for selection, some interesting conclusions may be reached. [163] Another practical point of the same nature arises from the great variability which these peas manifest in plant- as well as seed-characters. Mr Hurst of Burbage tells me that in _e.g._ _William the First_, a pea very variable in seed-characters also, tall plants may be so common that they have to be rogued out even when the variety is grown for the vegetable market, and that the same is true of several such varieties. It seems by no means improbable that it is by such roguing that the unstable mosaic or blend-form is preserved. In a thoroughly stable variety such as _Ne Plus Ultra_ roguing is hardly necessary even for the seed-market.