50 BULLETIN 1164, U. S. DEPARTMENT OF AGRICULTURE. 
afford conclusive cadence of linkage. Moreover, all or nearlv all of 
the characters involve more than one factor, and t is h%hh prohable 
that the several factors for a given character are distributed fn differ < 
SatioiTuthShiri,?- 34) i, Detomi , atw » of ^ actuil^Xg; 
mtuTof Sie *ar^tl W ° U ' d "*" l ° be PreCluded b >' ** ™»lA 
PRACTICAL SIGNIFICANCE OF THE RESULTS. 
Many unsuccessful attempts have been made bv practical breeders 
to obtain a superior type by hybridizing upland wiSi Egyptkn ol 
with sea-island cotton Combination of the large upland ffi with 
the long silky fiber and the relatively smooth seeds of the EbvS 
wnfdnt ^ Ti?" 8 , 1S an ^active goal, but one wh ch doffis 
« ould be very difficu t to reach. The difficulty of obtaining and fixing 
desirable recombinations from hybrids of upland with E^ ptian of 
sea-island cottons has been discussed by Cook (13 pn 6<Mi4) lu 
KSSJ}* ° ne ° f the m ° St ™ obstaclesSs'co^rence'of W th°e 
in Tl!! ^tf, 1 ^^. Presented concerning the correlations of characters 
£ l k Holdon-P,ma hybrid indicate that freedom of recombination 
L,f C It m ^ Upkl ? d an , d E S>'P tian cotton is m«ch 1 ™ t "i ed 
X i : j s »PP°?ed and evidence is also given that even as ear v as 
the hird generation the new combination! may be reladvcy stable 
mdiv H ° ^ ^ I* Sh ° Uk i b , e P° lnted out that " ™* ^e of tfe 215 F 
individuals and not one of the F 3 progenies showed any indication 
varietTT, a n gnCultural T"' com Pa™We to that of either ptn" 
Kni J Vl^ ne , ar a PP r oach to a recombination of upland size of 
dttZt Tn? t U f \° { fibe J and rdfttive ^oothness P of seeds was 
detected. The briefest consideration will show that the chance of 
L U ^rem™ e inatl ° n ^^ " tMs Sma " F > W«C»SS* 
Leaving all other characters out of account for the moment let 
us assume tha the breeder's goal would be attained f T boll 
diameter of Holdon could be combined witl, the length of fiber an 
relative smoothness of seeds of Pima. It may be inferred fromTheF 
frequency distributions for these characters (figs. 34. 3S and 41) that 
ors eXP On The ?n ^ ° ^^"P™ » ot less than thre fac- 
tors. On the theory of probability onlv one in 262,144 individuals 
homozygous for the nine factors involved, could be expected in F and 
to have a reasonable chance of getting the one honiozvgousmdi vidua 
at least three times this number of plants would be rfou red Even 
were it practicable to grow so large an F, population, discovery of the 
one desirable plant would be exceedingly unlikelv, especiallv as close 
examination ,s required to bring out thefiber anu seScharacters 
Emerson and East (17, p. 117) have pointed out, however that 
the chance of obtaining by hybridization a desired recombination 
involving numerous factors is not so remote as would appear from cT 
culation of the probability in the'second generation. P |b"pro£hie 
