Demonstration of the importance of sample size in confidence in data									page 27
23 Jan 1995, revised 2 Jan 1996, 5 Jan '98, 2 Jan 03
Cross:	AABB x aabb produces F-1.Then, AaBb x aabb test cross to produce progeny:
Null hypothesis:		There is no linkage between A and B, thus 1:1:1:1 ratio.
I.	Cross is performed with sample size of 300 progeny.
	phenotype	observed	expected	dev'tn	sqr of dev	sqr dev/exp
	AB	78	75	3	9	0.12
	Ab	74	75	-1	1	0.0133333
	aB	63	75	-12	144	1.92
	ab	85	75	10	100	1.3333333
	Total:	300			X sqrd:	3.3866667
					d.f. = 3	P = 0.1 - 0.5
	Thus, expect deviations as great as these 10% of the time.
	Therefore, we conclude that the markers have not shown linkage.
II.	Cross is performed with sample size of 3000 progeny.
	phenotype	observed	expected	dev'tn	sqr of dev	sqr dev/exp
	AB	780	750	30	900	1.2
	Ab	740	750	-10	100	0.1333333
	aB	630	750	-120	14400	19.2
	ab	850	750	100	10000	13.333333
	Total:	3000			X sqrd:	33.866667
					d.f. = 3	P.= > 0.005
	Thus, these deviations are expected randomly less than 0.5% of the time.
	Therefore, we conclude that data would deviate this much from the expected ratios
	only 0.5 % of the time by random variation if they are not unlinked, and the
	null hypothesis is rejected.
III.	Linkage:
		recombinant types:		Ab	740
				aB	630
		Total recombinants:			1370
		Total progeny:			3000
	Recombinant/total progeny:		0.4566667	=
45.7 map units between A and B.