頁籤選單縮合
題名 | 臺灣主要農作物取樣單位之初步研究=Preliminary Results from Study on the Sampling Units for Farm Crop Trials in Taiwan |
---|---|
作者 | 汪厥明; 葉樹藩; | 書刊名 | 中華農學會報 |
卷期 | 44 民52.12 |
頁次 | 頁9-17 |
關鍵詞 | 臺灣; 農作物; 取樣單位; |
語文 | 中文(Chinese) |
英文摘要 | The main aim of research works is to find out the precise and feasible methods of determing the size of sampling unit that can be applied in any experimental field of the same nature with special reference to the normality, efficiency, representativeness and economics of the sampling units. Then, we apply these methods in finding out the optimum size of field plots for purposes of farm crop trials. The procedures of normality test, maximum curvature and FAIRFIELD SMITH'S variance law are used in determining the size of sampling units. The Smith's original method, four-stage sampling method due to KOCH and RIGNEY and that of unbiased estimation with asymptotically minimum variance are employed to estimate the parameters, the coefficients of soil heterogeneity, in variance law. For evaluating the applicability, we first, review the theories underlying all methods involved herein, and then applied them to the data of uniformity trials, variety and treatment comparison, etc. with rice, sweet potatoes, spring soybean, tobacco and flax. From the reseach works above, it follows that the normality test is the good fundamental method of determining the size of sampling unit-the field plot-both for the theoretical and practical purposes, in any kind of field experiments whereas the maximum curvature method can only be utilized in finding out the suitable number of basic unit plots per sampling plot, i.e. within one field plot, and the variance law method in determining the economized efficient number of basic unit within a certain plot and also, in deciding the economized efficient size of the block. From the practical point of view, it is the best way to employ the normality test in determining the length of plot and the maximum curvature method for the number of rows per field plot. If the size of a single row and the number of rows per plot are ever able to be fixed, the size of plot will be automatically decided. A few words about the asymptotically minimum variance method due to HATHEWAY and WILLIAMS should be added that by the present researchers' experience, apart from the laborious calculation concerned herein, the practical advantage of applying this method, in spite of its relative soundness in theory, is not realized so much as expected by the two authors, because, so far as our experimental data are concerned, by their method, bw, the estimate of coefficient of heterogeneity thus obtained may turn out to be larger than one unit or otherwise, smaller than zero, and , besides, its estimated variance, v(bw), is often larger than the other methods involved in these researches. The suitable plot size for each farm crop trials worked out by us with the above methods is as follows: 1. Rice: The field plot should be at least 4.67 meters long for a single row plot, and it can be extended to a length of 9.33 without troubles. The optimum number of rows per plot lies somewhere between three and six rows, and appropriate size of plot ranges from 1.24 to 22.28 sq. meters. 2. Sweet potatoes: The plots of 4~12 meters long and four rows wide with an area of 16 to 48 sq. meters are suitable for field experiments. 3. Spring soybean: The optimum plot size is found to be 4~8 meters long and 3~6 rows wide with an area of 5.4~21.6 sq. meters. 4. Tobacco: The optimum plot size is 17.0~40.8 Tai-Chi long and 3~4 rows wide with an area of 165.2~522.24 (Tai-chi)2. 5. Flax: The optimum plot size is 8~16 meters long and 4~6 rows wide with an area of 3.6~21.6 sq. meters. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。