Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
COMPOSITION OF CORN IN THE UNITED STATES pies on which it was possible to determine the other minerals and vitamins varied consider- ably and was limited in many instances. Insofar as was possible, the regional allocation propor- tions were maintained in selecting the limited samples. The number of samples analyzed for each state for each year is shown in Figure 1. These numbers do not include samples for which calorie content alone was obtained. RESULTS Presentation of Data In order to have the results of both 1946 and 1947 analyses in one publication, the 1946 re- sults published in Report No. 1 are included in Table IA, and certain 1946 data not available until after the report was issued in Table IB. The results of the chemical analyses of the 1947 crop are shown in Table 1C. The means, the 95 per cent fiducial limits, and the number of samples analyzed for year and region are given in Tables IIA and IIB. The same estimates are given for the entire United States, as well as the standard deviations and the coefficients of variability. These standard deviations and co- efficients of variability are also applicable to any region. For each analysis, if it is stated that the population mean lies within the interval, sample mean ±L, the probability of being wrong is only 5 per cent. This may be said in another way: If it is stated that the true mean value lies within the confidence limits, the risk of being wrong is 5 per cent, or conversely, the chance of being correct is 95 per cent. The data for the two years are presented separately. For each nutrient, the number of samples analyzed is indicated. Variability of the Data The standard errors of the means are not given but may be easily computed, either on the absolute basis or as coefficients of varia- tion by dividing the standard deviations or the coefficients of variation, respectively, by the square roots of the numbers of samples. In general, the data for the two years are in - good agreement as regards both the means and the standard deviations. The only excep- tion is copper, where both the mean and the standard deviation are relatively larger in 1946 than in 1947, but only 42 samples were ana- lyzed in 1946 as compared with 197 in 1947. In several other instances the differences between means for the two years are statisti- cally significant. The nitrogen-free extract con- tent was significantly lower in 1947 than in 1946, and the contents of protein, fiber, ash, and some of the minerals and vitamins were significantly higher in 1947 than in 1946. Most of these differences may be too small to be of practical importance, but they seem to be sat- isfactorily explained by the fact that much of the corn in 1947 was soft or immature. The 1946 data are no doubt better to use as repre- sentative of commercial corn, at least in nor- mal years. Notable are the small coefficients of varia- tion for gross energy and the proximate nutri- ents. The low coefficient of variation for mag- nesium is striking as compared with those for most of the minerals, while that of fluorine is decidedly larger than that of the other min- erals. The coefficients of variability for thia- mine, niacin, riboflavin, and pantothenic and folic acids are not very large as compared with that for carotene. Locational Variations in Composition The relative importances of region and state- within-region variations in composition were studied by means of analysis of variance. Be- cause of insufficient data this analysis was not made for folic acid. The results are summarized in Table III. Some statistically significant regional and state-within-region effects were found. For most nutrients, however, these effects ac- counted for only a small portion of the total variance and are thus of questionable practical importance. Even where the regional and state variations were moderate or large, they may
COMPOSITION OF CORN IN THE UNITED STATES be of little practical nutritional importance. Thus copper, niacin, potassium, chlorine, and iron show moderate or large regional and state variations, but corn is not an important source of supply for these nutrients. Calories, nitro- gen-free extract, and fat show moderate re- gional and state variation and are supplied in important amounts by corn. However, the total variability over the nation is so small (see Table II) that the variance attributable to region or state would seem of little concern. Although of doubtful importance from the point of view of the practical nutritive value of corn, the magnitude and pattern of regional and state differences are of interest from the standpoint of nutrient composition of feeds in general. Although the size of the surveys was too limited to obtain a clear picture of the lo- cational variations in nutritive value, certain findings seem worthy of brief mention. Protein content was lower than average in a region approximately denned by the states of Illinois, Indiana, Ohio, Kentucky, West Vir- ginia, and Tennessee. With the exception of West Virginia, these states were sampled rela- tively heavily. This low-protein region ap- peared to extend southward and southwest- ward to include Texas, Arkansas, Louisiana, Mississippi, Alabama, and Georgia. With the exception of Texas, however, the sampling rates in these states were relatively low. Corn of about average protein content was obtained from two regions. The first of these was defined roughly by Minnesota, Iowa, and Missouri, where the sampling rate was heavy. New York and Pennsylvania formed the second region, but these two states were not sampled heavily. Corn of above average protein content was obtained in three regions: (a) North and South Dakota, Nebraska, Kansas, and Oklahoma; (b) Wisconsin and Michigan; and (c) the At- lantic seaboard states from New Jersey to South Carolina inclusive. With the exception of Nebraska and Wisconsin the sampling rates in these states were relatively light. In general, regions of similar composition did not neces- sarily correspond to climatic regions. The lowest fat content was observed in In- diana, Ohio, Kentucky, and West Virginia. This is roughly the lowest protein region, ex- cept that Illinois and Tennessee are not in- cluded. The southern states which appeared to be in the low-protein group did not fall in the low-fat group. Instead, it appeared that the southern states in general, including Texas and Oklahoma but excluding Virginia, Tennes- see and Kentucky, produce corn of higher than average fat content. The remainder of the states in the East and Midwest, which pro- duced corn of average or high protein content, in general showed average fat content although a few showed higher than average fat values. Location and variety (including color) were partially confounded in this study; hence the regional differences in fat content are possibly due to variety and color. With respect to the other nutrients studied there were often rather marked differences be- tween states, but in general no region having characteristic values could be discovered. One exception was iron content. The southern states in general showed low iron values and the mid- western, northern, and eastern states showed generally high values. There were exceptions, however. For most of the nutrients the num- ber of analyses was too small to warrant at- tempts to define regions as was done for protein. Effects of Soil Type on Proximate and Mineral Composition The results of the locational analyses sug- gested that perhaps soil type was a factor affecting the composition of corn. To supple- ment the locational analysis, therefore, the samples were classified by Great Soil Groups and by soil associations. Analyses of variance were then conducted on the proximate nutri- ents and the minerals, it being considered that these would show greater effects from soil than would the vitamins. The results are summar- ized in Table IV. As was true with location, soil classes were in general only slightly associated with varia-
COMPOSITION OF CORN IN THE UNITED STATES tions in the composition of corn. Fat content was the only factor that appeared to be asso- ciated with differences in Soil Groups or soil associations in any marked way, but there was a suggestion that protein, potassium, and iron may bear a relation to the Great Soil Group classification. The minerals, which might be expected to vary considerably with soil classes, were in general for the two years associated only slightly with them, although soil classifi- cations were significant in one of the years in several instances. Differences between White and Yellow Corn Analyses of variance were conducted to as- certain whether white and yellow corn differed in nutrient composition in any way other than carotene content. All of the samples of white corn were open-pollinated, but the yellow sam- ples included both hybrid and open-pollinated varieties. The data were insufficient to make comparisons on cobalt and folic acid in 1946 and on calories, potassium, magnesium, so- dium, chlorine, fluorine, thiamine, niacin, pan- tothenic acid, and folic acid in 1947. For none of the nutrients was there a signifi- cant difference between white and yellow corn in 1946, and in 1947 significant differences existed for fat and iron only. In both years the fat content of white corn was higher and the iron content lower than in yellow corn. In gen- eral, however, it appeared that the only practi- cal nutrient diffarence between white and yel- low corn b carotene. Varietal Variations in Composition Because all varieties of white corn were des- ignated simply as open-pollinated, varietal variation was studied only on yellow corn. Analyses of variance were conducted for each year separately, and only those varieties were included for which there were data on two or more samples, nearly all of which were hybrids, in the year being analyzed. Because of insuffi- cient data, determinations of calcium in 1946, calories in 1947, and sodium, chlorine, fluorine, and folic acid in either year were not made. Location effects were disregarded in making these analyses because of the computational difficulties involved in adjusting for them. The results are summarized in Table V. In few instances were varietal effects statistically significant in either year, and only for fat were they significant in both years. In general, varietal differences accounted for a small por- tion of the total variation in nutrient composi- tion. It appears that, with the possible excep- tion of fat, differences in nutrient composition among varieties grown in practice are not of practical nutritional importance. It is not in- tended to imply that nutritionally improved varieties cannot be developed by proper breed- ing and selection, but only that the more im- portant commercial varieties appear to be genetically rather uniform with respect to fac- tors determining nutritive value. Correlations between Nutrients The determination of certain micronutrient elements and vitamins is very difficult and te- dious. If it were discovered that high correla- tions existed between various nutrients in certain feedstuffs, difficult-to-determine nutri- ent values could sometimes be predicted from easily determined nutrient values. Such a dis- covery could possibly eliminate the necessity for conducting many difficult determinations and could be of practical importance in nutri- tion as well as in sampling feeds. In this study an effort was made to discover the existence of high correlations between vari- ous nutrients contained in corn, and over 200 correlation coefficients were computed for each of the two years. Simple correlation coefficients were computed for all pairs of nutrients for which there were sufficient data. Data were insufficient to compute correlations of folic acid with all other nutrients in both years, calories with all other nutrients hi 1947, and fluorine with potassium, magnesium, sodium, and chlo- rine in 1947. Some correlations were found to be significant in each year, but the results for the two years were not consistent. The signifi- cant correlation coefficients were generally
