June 1983, and (2) the availability of both an adjacent pond receiving the discharge and a more distant pond in the same drainage (with polygonal trough or surface connections to the adjacent pond).

Control and experimental ponds were selected on the basis of similarity in physical characteristics (including depth and temperature) and chemical characteristics (especially salinity).

We selected control ponds in areas with no previous development of drill pads within several miles. Figure 2 shows the location of sample and control sites. Three drill sites each were chosen in the Atlantic Richfield Company Alaska Inc (ARCO) (drill sites 6, 16, and 18) and in the Standard Alaska Production Company (SAPC) (drill sites A, E, and N) portions of the Prudhoe oil field. All adjacent ponds were within 25 meters of the gravel pad and were the direct receiving waters for the discharges from the reserve pits in June. Three of these ponds (at drill sites 16, 18, and A) were in contact with the gravel pad; two (drill sites 18 and A) had additional gravel contact and were impounded on at least one additional side. A fourth receiving pond (drill site E) was not in contact with the pad, but had an extensive gravel bottom, as a result of gravel-contaminated snow being bulldozed into the pond. Discharges resulted from overtopping of a dike at drill site 6 and truck spraying of the pad at site 18, rather than from the direct pumping of effluent into the tundra, as at other sites.

The distant ponds were 35 to 115 m from the adjacent ponds (average distance, 93 m). They were connected to the adjacent ponds bv polygonal

troughs. At the time of discharge, surface sheet flow also probably occurred between adjacent and distant ponds. Two of the control ponds were within the designated area off Oliktok Road shown in Figure 2. The

remaining control site was north of drill site E.

Hereinafter we designate sample sites by the drill sites named above, and by treatment, where reserve pits are treatment 4; adjacent receiving ponds, treatment 3; distant ponds, treatment 2; and control ponds, treatment 1. Treatment is used strictly in a statistical sense to

distinguish sample groups that are compared with each other, rather than in ■ biological sense to refer to a prescribed experimental treatment.

MATERIALS AND METHODS

Preliminary sampling began on June 6, 1983. We visited accessible production pads in the Prudhoe Bay oil field to determine which reserve pits were being dewatered and to quantify the number of pits with visible oil sheens and piles of drill cuttings. Grab samples were taken of discharge water for analysis of heavy metals and for hydrocarbon analysis. These samples and samples taken later in the summer are listed in Appendix Table Temperature, pH, and conductivity were also measured in the field during the June sampling. A HACH Mini-Digital pH Meter calibrated with buffer solutions was used to determine pH, and the remaining variables were measured with a YSI Model 33 SCT Meter and probe. The accuracy of conductivity measurements was checked with a conductivity standard.

A-1.

We followed methods recommended by APHA et al. (1981) in collecting water samples: both the preliminary (June) samples and later samples for heavy metal analyses were collected in laboratory-cleaned, 1-L polyethylene bottles and preserved with analytical-grade concentrated nitric acid; hydrocarbon samples were collected in 1-L, laboratory-cleaned glass bottlesand preserved with analytical-grade sulphuric acid. Samples were stored for several days at 0-4°C in coolers furnished with blue ice and then

refrigerated. Samples were subsequently shipped in coolers with blue ice to

the U.S. Fish and Wildlife Service Patuxent National Wildlife Research

Center (Laurel, Maryland) for analysis.

Six sample sites chosen in June were sampled again beginning July 27, 1983, about month after the last discharges from the pits onto the tundra. Samples were taken to be analyzed for heavy metals and hydrocarbons from the previous point of discharge in each adjacent pond. Temperature, pH, and conductivity were measured as before, in addition to dissolved oxygen, hardness, alkalinity, turbidity, water depth, and sediment depth. A HACH AL-36B Titration Kit was used to determine dissolved oxygen, hardness, and alkalinity. To check the strength of the phenylarsine oxide titrant for the dissolved oxygen test, we used an iodide-iodate standard solution equivalent to 10 mg/L dissolved oxygen. An HF Instruments Model DRT-15 nephelometric turbidimeter was used to measure turbidity, after calibration with secondary standards. Water and sediment depths were measured by probing a wooden dowel first to the water/sediment interface and then to the bottom of the soft sediment layer. Characteristics measured at each site are reported in Appendix Table B-1.

Identical samples for chemical analysis were collected at the reserve pit, at adjacent ponds that had earlier received the pit discharges, at the distant ponds within the same drainage, and at the control ponds.

Invertebrate samples were collected with a sweep net (30.5 x 7.6 cm)

having 7.9 meshes/cm, according to procedures used by Weller (1972); Howard

(1974); and Abraham (1975). About 1 was sampled for macroinvertebrates (Figure 3). Free-swimming, planktonic, and epibenthic invertebrates were all sampled by this method. Two complete sweeps were made from one location within the sample pond and captured organisms from each sweep were transferred to separate white enamel sorting trays for examination. If the samples appeared to be roughly similar, they were preserved together as one sample in 70% ethyl alcohol. If the samples did not appear to be similar, we discarded both to avoid bias as a result of making a sweep through an occasional swarm of Daphnia or copepods. This process was repeated at three locations so the three combined replicates from each pond represent six total sweeps for the site. This sampling was repeated at each of the six pits, six adjacent ponds, and six distant ponds, and at the three control ponds. Although the method was only semiquantitative, it enabled comparisons between treatments without introducing an even greater bias that can be incurred when the distribution of highly clustered populations is not specifically evaluated before quantitative sampling (Pielou 1978).

Except for nematodes, preserved invertebrates were sorted and identified to order, suborder or family. The various taxa used to determine taxonomic richness (1.e., total number of taxa) and taxonomic diversity

(Shannon-Wiener H' diversity), together with code designations for taxa, are shown in Appendix Table C-1.

The following methods were used in the analysis of metals samples: hydride generation atomic absorption (arsenic and selenium); cold vapor atomic absorption (mercury); and flame atomic absorption (aluminum, lead,