BOOK XVII. XXIV. 105-108 



is an old one, as in that case the sHp must be stouter. 

 A further point is that sHps that are going to be 

 grafted must be pregnant, that is, sweUing with bud- 

 formations, and in expectation of giving birth in that 

 year, and they must be at all events two years old, 

 and not thinner than the Httle finger. But grafts are 

 also insex*ted the other way round " when the in- 

 tention is for them not to grow so long but to spread 

 out. Before all things it will be serviceable for them 

 to have buds and to be glossy, as nothing shabby or 

 shrivelled anywhere will gratify one's hopes. The 

 pith of the siip grafted should be put touching the 

 place * in the mother tree where the wood and the bark 

 meet, for that is more satisfactory than to place it 

 level with the bark outside. The process of giving a 

 point to the sHp for grafting must not strip the pith 

 quite bare, but only make it visible through a narrow 

 aperture ; the point must slope ofF in an even wedge 

 not more than three inches long, which is most easily 

 achieved by dipping the sHp in water when paring 

 it. It must not be exposed to wind while it is being 

 pointed. The bark must not be allowed to become 

 separated from the wood in either the graft or the 

 trunk. The graft must be pressed right down to 

 where its bark begins, but it must not be forced out 

 of shape while it is being pressed home, nor have its 

 bark folded back in wrinkles. Consequently shoots 

 dripping with sap should not be used for grafting, no 

 more, I swear, than ones that are dry, because in 

 the former case excess of moisture causes the bark 

 to sHp, while in the latter owing to defective 

 vitaHty it makes no moisture and does not incorporate 

 with the trunk. Moreover there is a reHgious rule 

 that a graft must be inserted while the moon is 



75 



