CHARACTERS IN AX UPLAND-EGYPTIAN COTTON HYBRID. 51 
recommended is to select in F, and later generations those individuals 
which show the nearest approach to the combination in view and, if 
there is no indication that the object may be attained by selection 
alone, to intercross the most promising individuals or families. 
Another method to be considered in such cases is back crossing with 
whichever of the parent varieties has the larger number of desirable 
characters or the characters which had proved most difficult of 
recovery in F,. . . 
The employment of one or all of these methods might bring the 
combination of upland size of boll with Pima or sea-island length of 
fiber and relative smoothness of seed within the range of practicability 
were these the only characters involved. But so far the problem 
has been stated in its simplest terms. Habit of growth and fertility 
of the plant and additional properties of the fiber, such an abun- 
dance, strength, and fineness, would have to be taken into account in 
breeding a type of cotton desirable from the agricultural and commer- 
cial points of view. It was observed that the few second-generation 
Holdon X Pima plants which produced very long fiber were defi- 
cient in either the abundance, 34 strength, or fineness of the fiber. 
Probably the most serious obstacle to success in recombining the 
desirable qualities of such diverse types of cotton is the marked 
tendencv to sterility manifested in the perjugate (second and later) 
generations. If the great decrease in fertility from the second to the 
third generation, noted in the case of Holdon X Pima, is a phenom- 
enon of general occurrence in hybrids of upland and Egyptian cottons, 
the possibility of developing in this way a satisfactory agricultural 
variety would seem to be very remote. 
Fortunately, the chances are much greater of obtaining desirable 
recombinations in crosses between more nearly related cottons. The 
results of an earlier investigation of a hybrid between two varieties 
of Egyptian cotton proved that segregation in F 2 is comparatively 
slight and that a cross of this nature does not result in the appear- 
ance of extraparental characters, abnormalities, and sterility, as is 
the case when upland and Egyptian cottons are crossed. The prac- 
tical breeding work with Egyptian cottons has afforded evidence 
that it may not be extremely difficult to combine the desired char- 
acters of two such varieties in a stable and uniform type. 35 
It is very clear from the results of this investigation that accidental 
cross-pollination of such different cottons as upland and Egyptian is 
sure to depreciate greatly the agricultural value of either type. The 
extensive segregation must result in a loss of that uniformity in the 
length and other properties of the fiber upon which its commercial 
value so largely deoends, while the partial or complete sterility of 
many of the segregates will necessarily reduce the yield. 
The recognition and definition of numerous characters in hybrids 
between upland and Egvptian cottons which are not expressed in 
either parent, some of these being of a very striking or abnormal 
description, has a direct application in the practical work of main- 
taining supplies of pure planting seed. It is evident that a stock of 
Egvptian cotton can not be regarded as uncontaminated by cross- 
