necessary for nut production and must be present in abundance as its
movement in the orchard is subject to the vagaries of the wind, and only
a small percentage of that in the air ever comes in contact with the
stigmas of the other varieties.

It is the purpose of this paper to indicate the value, insofar as it may
be estimated from the available data, of the different varietal crosses
in obtaining desirable filbert hybrids. Table 1 contains a list of
crosses made, the number of seedlings raised, and the percentage of
these which were of sufficient merit to be retained for further study.
The percentage of seedlings propagated indicates even more definitely
which crosses are of the greatest value in producing superior seedlings
as only the outstanding seedlings were propagated for a second test.
Selections included in Table 1 are there by virtue of their all-around
merit.

Crosses between Rush and Littlepage and Rush and Winkler produced
nothing of value. The populations were small, but other equally small
populations from other crosses produced seedlings of value. The
inter-crossing of selections of _Corylus americana_ does not appear to
be a promising line of attack in filbert breeding where hybrids with _C.
Avellana_ will thrive.

Rush and Barcelona were each used as seed parents in crosses with the
same eight varieties. In the crosses involving Rush 1,232 seedlings were
produced and of these 39, or 3.2%, were good enough to propagate. Of the
306 seedlings raised from the same varieties combined with Barcelona
only 4, or 1.3% were worth propagating. None of these Barcelona
seedlings are among the best. Under the conditions of the experiment it
would seem that Rush is much superior to Barcelona as a parent in
crosses with varieties of _Corylus Avellana_.

The cross between Kentish Cob and Cosford failed to produce any
seedlings of outstanding merit.

In considering the productiveness and hardiness of the catkins of the
seedlings resulting from the different crosses the data have been
assembled in Tables 2 to 5, each table containing the summarized records
for different plantings. These plantings were started at different times
and the records are not directly comparable as they are for different
years and varying lengths of time. In Table 1 the total number of
seedlings is given, but in Table 2 to 5 only the data for the selections
are used. Records for the selections are available for several years,
whereas th