SCHERBA and GALLUCCI: SYSTEMATIC SAMPLING OF INFAUNA 



placed parallel to the water, do not show the 

 gradient. This is probably a consequence of the 

 short distance between strata. 



Field Procedures 



Within each region (Figure 1), a rectangular 

 study area was defined, measuring 95 m in length 

 parallel to the waterline, and approximately 7 m 

 wide, perpendicular to the waterline. Two parallel 

 strata, approximately 2.5 m apart, were placed 

 within each area, parallel to the waterline. The 

 stratum at the highest tidal level was designated 

 stratum 1 while the lower stratum was designated 

 stratum 2. Stratum 1 within the areas was located 

 at -1.4, -1.2, -1.1, and -1.2 feet in areas 1, 2, 3, and 4 

 respectively; while stratum 2 in those same areas 

 was at -1.5, -1.7, -1.5, and -1.6 feet."' The study 

 areas were numbered one (1) to four (4) (south to 

 north) and defined by stakes marked with 

 fluorescent tape for night identification. 



It is necessary to test the homogeneity of 

 sediment composition within a region if the areas 

 are to be considered representative. This test was 

 accomplished by dividing each area into two 

 subareas,*' separated by 5 m, and denoted as A (for 

 the northmost subarea) and B (for the southmost). 

 Each subarea contained about 448 sampling units. 

 Two samples were then collected on each stratum, 

 from each subarea, using a systematic sampling 

 plan. 



Each subarea was considered to contain sepa- 

 rate populations, and the two population Kol- 

 mogorov-Smirnov procedure with n = 4 (Conover 

 1971) was applied to the data collected. This use of 

 both subareas was carried out only for the winter 

 sampling. Winter sampling of the subareas was 

 done on: 8 January 1974 (lA, IB), 9 January 1974 

 (3A, 3B), 2 February 1975 (2A, 2B), and 3 February 

 1974 (4A, 4B). 



Spring and summer sampling was conducted 

 only in subareas IB, 2A, 3A, and 4A as follows: 24 

 May 1974 (IB, 3A), 21 June 1974 (2A, 4A), 19 July 

 1974 (IB, 4A), and 16 August 1974 (2A, 3A). Each 

 stratum in these four subareas was independently 

 sampled during these two seasons with n = 4 on 

 each stratum. 



All samples were collected using a thick-walled, 

 cylindrical corer made of polyvinyl chloride pipe, 



10 cm inside diameter and 18 cm long. The corer 

 was pressed into the sediment to 18 cm, and its 

 contents removed by hand, placed in a labeled 

 plastic bag, and taken to the laboratory. Each 

 sample was passed through a 1-mm sieve, and the 

 contents retained by the sieve were sorted twice 

 by eye to remove all bivalves and polychaetes (the 

 only members of the macrofauna identified). 

 These organisms were placed in 80% ethanol and 

 8% formaldehyde, respectively, for later 

 identification. Only the common bivalves and 

 polychaetes were identified to genus and species. 

 The sediment portion of each sample was dried at 

 100°C for approximately 4 h. The method used to 

 quantify the particulate properties of the 

 sediment was the percentage composition by 

 weight of selected sediment grain sizes. A me- 

 chanical shaker was used to pass the sediment 

 portion of each sample through a series of 

 Wentworth sieves (1.981, 0.495, 0.246, 0.124, 0.063 

 mm). The contents of each sieve were weighed and 

 recorded as percentage of the total weight of that 

 sample. 



Statistical Procedures 



Estimates of the variances of the sample means, 

 obtained from SSS were approximated by the 

 estimate of the variance of the sample mean from 

 a simple random sample (see Cochran 1963), i.e., by 

 using 



var 



where 





*Tidal heights are reported in feet to conform with U.S. Coast 

 and Geodetic Survey Tide Tables. 

 ®We thank A. R. Sen for this suggestion. 



The rational for this approximation is discussed 



later. 



Two statistical tests were used to quantify the 

 sampling results. The K sample Kolmogorov- 

 Smirnov(K-S)testwitha = 0.10, using the Tg test 

 statistic (Birnbaum and Hall 1960; Conover 1971), 

 was used to test hypotheses about variation in 

 sediment composition. The chi-square test for 

 several multinomials with a = 0.05 (Conover 1971) 

 was used to test hypotheses about variation in 

 bivalve and polychaete community structure. 



In the within-area sediment homogeneity test 

 empirical distribution functions were constructed 

 for each subarea. The K-S test (a = 0.10) was then 

 used to test the null hypothesis (Ho) of equality of 



939 



