Nov. 26, 1917 
Field Technic in Rod Row Tests 
415 
Twenty plots were used for the Farm Crops test and 59 for the Plant 
Breeding work. There appears to be very little difference in variability as 
determined by the Farm Crops or Plant Breeding results, the Plant Breed¬ 
ing coefficients of variability for the respective replications being only 
slightly larger than those which were obtained for the Farm Crops test. 
In the Farm Crops test there is a large reduction in the standard 
deviation for two replications as compared with a single replication. A 
slight reduction in variability for three as compared with two replica¬ 
tions is apparent. Three or four replications appear to be about equally 
reliable. There is a large reduction in the standard deviation due to six 
replications as compared with three. For the central row tests the 
difference in the standard deviation for yield of six replications as com¬ 
pared with three is 1.80±0.884. This is slightly more than two times 
the probable error. 
The comparison of 3-row plots and central rows shows little difference 
for any number of replications except nine, and such differences as are 
obtained for this number of replications are in favor of the use of central 
rows as compared with 3-row plots. 
For the Plant Breeding test there is little difference between the use of 
3-row plots and central rows for any number of replications below nine. 
There is an apparent advantage in favor of the use of 3-row plots for 
nine replications. As this result is exactly the reverse of the Farm 
Crops result for nine tests, it would indicate that these results are chance 
fluctuations. 
For the central rows three replications appear to be as reliable as any 
number below nine. There is some advantage in the 3-row plot tests for 
six as compared with three replications. This is, however, of doubtful 
significance, the difference in the standard deviations being 0.99 ±0.541. 
Nine replications appear to be as reliable as nine replications for the 
Farm Crops work. It is of interest to note that three replications reduce 
the error about 50 per cent as compared with a single test. 
The results for replications in barley are given in Table XIII. 
Table; XIII .—Value of replication of central rows and 3-row plots of Manchuria barley 
(Minnesota 105) used as a check in the Farm Crops and Plant Breeding tests 
Designation of test. 
Method of test. 
Number 
of tests. 
Mean. 
Standard 
deviation. 
Coefficient of 
variability. 
Farm Crops variety test.. 
Single central rows. 
15 
3 a.o±o. 331 
i.9©±o. 234 
5 - 94 ±o. 731 
Do. 
2 replications. 
7 
32.od= .428 
32* oi .317 
32. o± .386 
32. oi .191 
i.68i .303 
i.o$± .224 
• 99± *273 
• 4 ©± • 135 
5 * 25 ± .946 
3. 28i .670 
3.09± .851 
1.25± .422 
Do.*. 
3 replications. 
5 
Do. 
4 replications. 
3 
Do. 
6 replications. 
2 
Do. 
3-row plots. 
15 
7 
3i-3± .298 
31. 2± .247 
31* 3± *284 
1. 7ii .211 
i.36± .245 
. 94± . 200 
5 * 4 <>i .672 
4 * 3 <S± * 7 ^ 
3*oii .642 
Do. 
2 replications. 
Do. 
3 replications. 
5 
Do. 
4 replications. 
3 
3i*3± .167 
• 43 ± .118 
i* 37 ± -377 
Do. 
6 replications... 
2 
3 t* 3 ± *095 
36.2 ± .949 
. 2oi .067 
4.22± .672 
.64± .216 
11.66i1.872 
Plant Breeding Nursery.. 
Single central rows. 
9 
Do.. 
2 replications. 
4 
36.6± .907 
2.69± *642 
7 * 3 S^i* 7 S 3 
Do. 
3 replications. 
3 
36. 2±1.050 
2. 7©± .743 
7.46i2.oS4 
Do. 
3-row plots.. 
9 
37 - o± .924 
37 -6± .853 
4 -n± *653 
2. 53 i .603 
11. nil. 784 
6. 73 i*. 6 os 
Do. 
2 replications. 
4 
Do. 
3 replications. 
3 
36.9± 1.102 
2.83* .779 
7.67i2.112 
15755°—!7 - 2 
