RELAPSE AND RETROGRESSION. 75 



and Gesnerias.. Un the other hand, variegated or 

 largely altered flowers of the simpler types rarely 

 occur even in cultivation. Hence we may, perhaps, 

 reasonably infer that great readiness to assume new 

 colours affords in itself a certain slight presumption 

 in favour of some previous colour modification. We 

 shall apply this hypothetical principle in the sequel to 

 sundry cases of yellow colouration in apparently high 

 families, as one among several tests by which we may 

 be aided in distinguishing retrograde from primitive 

 yellowness. 



This seems also the proper place to consider the 

 proofs of the position already advanced, that new 

 colours make their appearance at the edge of the 

 petal, and gradually work their way inward. Four 

 such proofs may be advanced. In the first place, 

 purely adventitious individual colours almost always 

 so appear. For example, the reddish tinge occa- 

 sionally observ^ed on many yellow flowers is usually at 

 the tip : so is the lilac tinge on certain white anemones, 

 and the pink tinge on many crucifers and umbellates. 

 In the second place, the slight blush which occurs 

 normally on flowers like the daisy, the apple-blossom, 

 and the blackberry, and which appears to be as yet 

 comparatively uninfluenced by insect selection, seeing 

 that it is deepest on the back of the petals, generally 

 occurs near the tip. The same peculiarity may be 

 observed also in several small CaryopJiyUaccce, Papi- 

 /ionaccic, and Ericacccc. In the third place, flowers 

 which open pink, like so many BoraginacciCf and then 

 become blue, remain always red at the base, and only 

 acquire the nev/ hue in the ex[)andcd limb of the corolla. 

 In tulips, Hydrangea^ Ric/uirdia, &c, like facts occur. 



