Sample size influences estimation of cotton dry matter, nitrogen accumulation, and lint yield1

Abstract

Accurate estimates of cotton (Gossypium hirsutum) dry matter accumulation and nitrogen content are important for both production and environmental reasons. One of the important factors in estimate accuracy is sample size. The objective of this investigation was to determine the cotton sample size necessary for acceptable estimates of cotton dry matter, lint yield, and shoot N per 100 kg of lint ratio (NLR) values. Three cotton cultivars (DeltaPine 90, DeltaPine 5415, and Stoneville 474) were planted on 13 May 1997 in an Eunola loamy sand (fine‐loamy, siliceous, thermic Aquic Hapludult) in 9.3‐m2 subplots of a split‐split plot design. Split plots were four sampling dates. Split‐split plots were four sampling techniques [a) four randomly selected plants (4RP), b) 0.3 meter of row (0.3‐m), c) one meter of row (1‐m), and d) two meters of row (2‐m)]. Each entire subplot was harvested on each sampling date after sampling by the four techniques. Shoot dry matter for the whole plot was 7.2 Mg ha‐1, and lint yield was 1.46 Mg ha‐1. Cotton shoot dry matter and NLRs were significantly overestimated by both the 4RP and 0.3‐m techniques, but not by the 1‐ and 2‐m techniques. The NLRs of cultivar subplots varied with cultivar from 9.1 to 11.4. The earliest maturing cultivar, DeltaPine 90, had the lowest NLR and the latest maturing cultivar, Stoneville 474, had the highest NLR. Accurate estimates of cotton dry matter accumulation and N content will likely require 1‐m samples, and 2‐m samples should further improve precision. The NLRs were similar to data (NLR <15) that suggest 1.6 Mg ha‐1 (3‐bale/acre) cotton lint yields can be achieved with less than 250 kg ha‐1 of shoot‐accumulated N.