CHAPTER  III



                      Hybridising and Selection.



 Perhaps  the  most fascinating of all branches of plant cultivation  is the

 production  of new forms in the  garden itself.  New plants of garden

 origin, as  distinct from  those  newly introduced  from other countries, are

 obtained in three ways t by  branch " sports,"  by selection among seed-

 raised plants, and  by  hybridisation.  So  far  as  trees" and  shrubs  are

 concerned, the first process is purely accidental, the  second frequently so.

    (i) Branch sports are abnormal shoots that occasionally appear on

 adult trees or shrubs and  are taken off and propagated by cuttings, buds,

 or grafts.   Many  of them  preserve their  abnormality indefinitely,  but

 others have a strong  tendency to  revert to the  normal type.   Nearly all

 variegated shrubs and trees, those with  deeply cut leaves, and those with

 double flowers originated as branch sports.

    (2) The production of new forms under cultivation by selection from

 seed has given to gardens some of their most  beautiful  plants; but in

 regard to  trees and shrubs  (trees especially), the intervals between the

 generations are too  long for the work to attract the ordinary man as a set

 purpose.   Most  new forms of seedling origin in gardens have originated

 as  chance  breaks, noticed by nurserymen or others among  batches of

 plants raised "to  furnish  ordinary  stock.  Most weeping,  fastigiate and

 dwarf trees  have originated  in  this way; also purple-leaved,  large- or

 small-leaved varieties, and forms  with richer-coloured or larger flowers.

 Like  branch sports,  they must be increased by vegetative parts—cuttings,

 grafts, etc.—and  propagated  in this way they show little or no tendency

 to  revert back to  the  normal  type.  Raised  from seed  they show a

 strong but  not a uniform tendency to  revert to the parent type;  thus

 often a small proportion come true  or even show the peculiarity of the

 form  to an increased extent;  a larger proportion are more or less inter¬

 mediate;  the remainder will  be  indistinguishable  from the type.  The

purple beech is an illustration; comparatively few of its  seedlings come

quite  true, they are mostly of an  ineffective  purplish green or  coppery

hue, but a number of purple beeches have been raised from seed, such as

        25