In the present study, a 120-h exposure to dispersed oil resulted in concentration-dependent mortality in C. finmarchicus during exposure and recovery. Furthermore, surviving females were monitored for reproduction ability during a 3-wk recovery period. The copepod groups exposed to the medium and high concentration of dispersed oil (1.8 mg L−1 oil and 16.5 mg L−1 oil, respectively) apparently displayed a short-term decrease in egg-laying rate (female−1 day−1) relative to the control and low-concentration groups, thus suggesting impaired or delayed reproduction ability. Following a period of recovery, however, the egg production of these groups resumed; and from around day 15 of the recovery period and onward, the calculated egg production rates in the high-exposure group significantly exceeded those of the controls. For the 2 lowest exposures, no significant differences in reproduction rates from controls were found from day 16 onward.
Postexposure reproduction recovery dynamics
A clear-cut, concentration-dependent depression on reproductive capacity of the exposed groups was demonstrated after terminating the oil exposure. The monitoring of nauplii in the recovery tanks started 5 d after the exposure ended, and the number of nauplii (female−1) in the high-exposure groups was still approximately one-tenth of that in the control. The calculated egg production (eggs female−1 day−1) compared with control was even lower in both the high- (p < 0.01) and medium-exposure (p < 0.05) groups (Figure 4A), and the swimming behavior of the females seemed affected (more sluggish). The data and observations are comparable with results from a study reported by Jensen et al. , in which decreased egg production and hatching success were observed in C. finmarchicus exposed to 10 nM and 100 nM of the PAH pyrene, respectively. Jensen and Carroll  also found concentration-dependent reduced hatching rates in Calanus glacialis Jaschnov exposed to North Sea crude oil water soluble fraction (consisting of 16 PAHs identified by the US Environmental Protection Agency, concentrations ranging 3.6–10.4 µg L−1). In contrast, in a study in which the calanoid copepod Centropages hamatus (Liljeborg) was exposed to 20 µg L−1 to 80 µg L−1 dispersed oil for 48 h to 64 h, Cowles and Remillard found no effect on egg production rates but did observe a clear concentration-dependent decrease in egg viability (hatching success) . No data for hatching success from the early recovery period in the present study are available, but the difference in egg production response between C. finmarchicus in the present study and C. hamatus in the referred study  may be related to different exposure conditions, oil composition, or species-specific tolerances.
As opposed to most previous studies, in the present study we were able to follow the recovery process over an extended period of time. As noted briefly above, clearly impaired C. finmarchicus females in the high- and medium-concentration groups gradually recovered; and after day 17 (11 d postexposure), the females exposed to the highest dispersion concentration had a higher calculated egg production than any of the other groups. This persisted throughout the remainder of the period (up to day 25; Figure 3). The calculated average egg numbers produced per female per day over the entire period was actually significantly higher in the highest exposed group than all the other groups (Figure 4C). This may be explained by a significant compensatory reproductive response in this group after the egg production was temporary depressed (see further discussion). Nevertheless, the total calculated offspring production in the tanks in which females recovered from the highest exposure concentration was still significantly lower than in the control recovery tanks (Figure 5). Hence, the increase in reproduction rate was not sufficient to compensate for the loss of females resulting from acute mortality and/or loss of fertility during exposure and recovery.
Although our argumentation thus far may be essentially valid, the discussion needs some refinement when the calculated egg production numbers from the recovery tanks are compared with the egg production data from the SFR test. For the control and low concentration groups, the calculated daily egg production never exceeded 20 and most of the time remained at a level less than one-fourth of the average daily production revealed from the SFR test (40–50 for all groups; Figure 6B). In addition there was no increase—or at most a weak increase—in the calculated egg production over the recovery period. For the medium-exposure groups, in contrast, the egg production was quite low at the start of the period (approximately 5 per female d−1) but increased steadily to approximately 20 at the end of the period. This tendency was even more pronounced for the highest exposed groups, where almost no eggs were produced in the start of the period (Figure 3D), but the production in the last part of the period approached the numbers from the SFR test (Figure 4B).
From the outset, we hypothesized that the control and low-exposure groups would demonstrate egg production capacity similar to the results from the SFR test, whereas the more exposed groups would show dose-dependent reduction in the egg production ability, at least initially. The low initial egg production calculated for the medium- and high-exposure groups seems to support this hypothesis, while the low egg production in the control and low-exposure groups apparently contradicts it. It should, however, be emphasized that the above considerations are based on calculated egg production and not direct egg counts, and the calculations may therefore be inaccurate. The calculations were based on the daily nauplii counts, female counts at the start and stop of the recovery period, and input parameters from the single female reproduction test and from reliable literature (see Materials and Methods). But nauplii mortality could not be accounted for, and the calculated egg production actually presupposed 0 mortality.
To counteract both food deficiency and possibly cannibalism, feed algae was added in concentrations that, according to Campbell et al. , should be enough for proper egg-laying and nauplii development. The concentration was verified through daily analyses by particle counting. Feeding in all recovery tanks was evident from the presence of fecal pellets on the bottom, but feeding activity could not be measured quantitatively in the applied system. Oil components have been shown to depress feeding activity in copepods [9, 35], and reduction in feeding into the first part of the recovery period could very well add to the low initial egg production in the highest exposed groups. However, the SFR test demonstrated egg-laying rates comparable with rates reported as normal for the species . Furthermore, because C. finmarchicus is an income breeder , the feeding ability was probably restored at this time of recovery. We are therefore apt to believe that food scarcity or inadequate feeding can hardly explain the peculiarities in the results.
Although the SFR test indicated normal feeding in all groups in the last part of the recovery period, the calculated egg production in the recovery tanks of the control and low- and medium-concentration groups did not parallel the much higher egg production in the SFR test, despite being fed the same diet. Only the highest exposed groups showed egg production rates similar to the SFR test. Although no irrefutable evidence is available, cannibalistic grazing on nauplii may be involved because both C. finmarchicus and relatives are well-recognized cannibals [37, 38]. For C. finmarchicus, Basedow and Tande  found very high nauplii clearing rates, averaging to 689 mL female−1 d−1 and in some cases up to 2 L female−1 d−1, with only a slight reduction throughout the range 0 L−1 to 20 nauplii L−1. They also found a linear relationship (measured up to 20 nauplii L−1) between nauplii abundance and predation rate (slope, y = 0.4338x at 6 °C). Correspondingly, Landry  reported nauplii predation rates in Calanus pacificus Brodsky that did not level off even at nauplii densities close to 200 L−1. Basedow and Tande  found no feeding saturation within the applied levels of algae feed, and with particular relevance to the present study, cannibalizing did not seem to be reduced, even at high feed algae concentrations.
Applied to the present study, and considering reproduction dynamics, the predation rate reported by Basedow and Tande  would almost triple the calculated egg production rate in the control and low-exposure groups. A certain further upward revision of the predation rate would be necessary, however, to increase the egg production rates to levels revealed from the SFR test. Compared to the study of Basedow and Tande , a higher temperature was applied for the present study (10 °C vs 6 °C), and the tanks were rather crowded (up to more than 4 females and 120 N1 + N2 nauplii L−1). A further increased predation pressure on the nauplii in the present experiment may therefore be expected.
As opposed to the other groups, the high egg-laying rates in the high-exposure groups in the last part of the recovery period were comparable to the SFR test data, pointing to low nauplii mortality. Also, application of the predation rate from Basedow and Tande  to the data from these groups returns unrealistically high egg-laying rates (significantly above SFR test data). The data are hence consistent with low or absent predation in these groups. Also, the high egg-laying rates recorded from the last part of the recovery period indicate sufficient feeding ability.
Based on the above reasoning, the results can be evaluated as follows: 1) egg production rate was acutely affected by the exposure in a dose-dependent manner; 2) during the recovery, a large portion of the exposed females recovered and resumed egg production; 3) the calculated egg production rate of females exposed to the highest concentration was close to actual (SFR test) numbers, possibly due to less nauplii grazing (fewer and/or less viable adults); and 4) the unexpectedly low egg-laying rates calculated for the controls and lower exposures may be explained largely by extensive nauplii grazing.
Results considering environmental risk and damage assessment
The main focus of the present study has been the fate of surviving C. finmarchicus after an acute oil exposure incident. The rationale behind this approach has been to better understand how endpoints reported from laboratory exposure studies relate to effects on population or ecosystem level. Laboratory studies are used extensively to evaluate environmental impact of pollution, and reported results add to the basic scientific data applied for risk and impact analysis as basis for legislation on, for example, offshore oil emissions and contingency planning.
Most laboratory exposure studies on copepods have so far reported on acute test results or, more generally, results obtained during and/or at termination of exposure. Reported endpoints include lethality [11, 14], reproductive effects , and recently transcriptional or metabolic markers [2, 10]. However ecologically relevant, none of these endpoints is easily applicable to calculating long-term impacts on the target species or its natural community without information on recovery ability and life strategies of the species. A population of a species with high recovery ability, high reproduction potential, and corresponding high natural mortality may be more apt to recover quickly after an acute exposure incident. On the other hand, this should not be taken as a general rule. Neither should the recovery ability under laboratory conditions negligently be assumed valid for field conditions without further considerations. Animals surviving short-time exposure apparently unharmed may suffer long-range injuries, and food-web interactions may corrupt the validity of the laboratory results. As an example, enhanced predation on exposed individuals due to temporary impaired mobility may add to population effect.
In the present study, the tested cohort of reproductive C. finmarchicus demonstrated a surprisingly high ability to recover from transient high, sublethal exposure to dispersed oil, and the effects of the oil exposure on reproductive output were dominated by acute mortality during and shortly after exposure. Furthermore, the majority of the surviving population eventually reproduced at a normal rate, even after exposures close to the acute median lethal concentration. If verified that the results adequately mirror field conditions, the results hence indicate that risk and impact analysis of oil emissions may be fundamentally based on dose–response relationships from acute toxicity tests with copepods, provided additional lethality during the early recovery period is included. Although the reproduction capacity may prove to recover quickly under field conditions, other potential long-range effects on ecological fitness may significantly add to population impact, such as increased predation due to exposure-related reduction in mobility or simply egg-laying delayed beyond the favorable algal bloom period.