156 LATHYKUS ODORATUS. C'h&f. V. 



down, in the same manner as bees ought to do ; and then pods 

 were generally formed, but from some cause often dropped off 

 afterwards. One of my sons caught an elephant sphinx-moth 

 whilst visiting the flowers of the sweet-pea, but this insect would 

 not depress the wing-petals and keel. On the other hand, I have 

 seen on one occasion hive-bees, and two or three occasions the 

 Megachile willughbiella in the act of depressing the keel; and 

 these bees had the under sides of their bodies thickly covered 

 with pollen, and could not thus fail to carry pollen from one 

 flower to the stigma of another. Why then do not the varieties 

 occasionally intercross, though this would not often happen, as 

 insects so rarely act in an efficient manner ? The fact cannot, as 

 it appears, be explained by the flowers being self-fertilised at a 

 very early age ; for although nectar is sometimes secreted and 

 pollen adheres to the viscid stigma before the flowers are fully 

 expanded, yet in five young flowers which were examined 

 by me the pollen-tubes were not exserted. "Whatever the cause 

 may be, we may conclude, that in England the varieties never or 

 very rarely intercross. But it does not follow from this, that 

 they would not be crossed by the aid of other and larger insects 

 in their native country, which in botanical works is said to be 

 the south of Europe and the East Indies. Accordingly I wrote 

 to Professor Delpino, in Florence, and he informs me " that it is 

 the fij^d opinion of gardeners there that the varieties do inter- 

 cross, and that they cannot be preserved pure unless they ara 

 sown separately." 



It follows also from the foregoing facts that the several varieties 

 of the sweet-pea must have propagated themselves in England 

 by self-fertilisation for very many generations, since the time 

 when each new variety first appeared. From the analogy of the 

 plants of Mimulus and Ipomcea, which had been self-fertilised 

 for several generations, and from trials previously made with 

 the common pea, which is in nearly the same state as the sweet- 

 pea, it appeared to me very improbable that a cross between the 

 individuals of the same variety would benefit the offspring. A 

 cross of this kind was therefore not tried, which I now regret. 

 But some flowers of the Painted Lady, castrated at an early 

 age, were fertilised with pollen from the Purple sweet-pea ; and 

 it should be remembered that these varieties differ in nothing 

 except in the colour of their flowers. The cross was manifestly 

 effectual (though only two seeds were obtained), as was shown 

 by the two seedlings, when they flowered, closely resembling 



