Biological Assay of Available Energy with Growing Chicks

I. Factors Affecting the Accuracy of the Estimates

Abstract

Influences of dietary protein level, length of experimental period, difference in breeds of chicks and seasonal variation were evaluated on the data obtained by biological assay of available energy in feedstuffs with growing chicks.
1. Effect of dietary protein level
Mathematical analyses of the relationship between body weight gain and dietary levels of energy and protein revealed that the protein level of the reference diets for biological assay of available energy with day-old and 4-week-old chicks should be kept about 23% and 21%, respectively. Small variation of dietary protein level was expected to have little influence on the result of the bioassay.
Three diets containing 19, 21 and 23% of protein with constant energy level were fed to chicks of 4-week-old chicks for 4 weeks in Expt. 509. Growth rate of chicks on these diet were almost identical, suggesting that the mathematical considerations is reliable. Therefore, available energy of feedstuffs containing protein can be estimated by directly replacing cornstarch in the reference diet, provided the increase of dietary protein level by this replacement is less than 2%.
2. Effect of length of experimental period
Available energy of test diets (x) can be estimated from the weight gain (y) of chicks on the diet with the equation (1), where a is constant, b regression coefficient, r correlation coefficient, S_yy and S_xx sum of squares of y and x, respectively. These figures are estimated with the data on reference diets. n and p in this equation are numbers of lots of chicks fed the reference and test diets, respectively. t_α(n-2) is t-value in t-table
x=(y-a)/b±t_α(n-2)•√S_yy(1-r²)/n-2•√1/n+1/p+(x-x)²/S_xx/b……(1)
with probability of α and (n-2) degrees of freedom. Therefore, the bigger the n, p and b are, and the closer to 1 the γ is, the smaller fiducial limit for the estimate x is expected.
In Table 3, length of experimental period, b, 95% fiducial limit for b, r and 95% fiducial limit for x were summarized. It is suggested that 4 weeks of experimental period is necessary to get satisfactory result of bioassay, when starting chicks are used. This is because of the smaller b and r, and therefore, larger fiducial limit for x is obtained with experimental period shorter than 4 weeks. Two weeks of experimental period were shown to be quite satisfactory with 4-week-old chicks. At the slight sacrifice of thee accuracy of the estimate, the period could be as short as 1 week.
Bioassay with starting chicks has an advantage of estimation without regard to the difference in feed intake of chicks, while bioassay with older chicks has a merit of a shorter experimental period.
3. Difference in breeds of chicks and seasonal variation
In Table 4, b value, i.e. the change of body weight gain per unit change of dietary energy level, of various pure- and crossbred chicks were compared. No significant difference in b value was observed among the different breeds. No consistent seasonal variation of b was obtained in various experiments conducted in different seasons.