Home | Print Page | Contact Us | Sign In
Harmonization W.G.: Chironomus methods development
Group HomeGroup HomeGroup PagesDirectory & Features Join Group
Share |

In November 2011, Dave Mount of the U.S.EPA in Duluth wrote a memo to the Office of Research and Development on suggested requirements and performance criteria for laboratories conducting sediment toxicity tests.  In regards to toxicity testing with C. dilutus, he had the following suggestions:


Laboratory Performance Criteria

While the freshwater sediment toxicity test methods (USEPA 2000; ASTM 2011) specifyminimum test acceptability criteria for individual tests, it is recognized that these criteria represent only minimally successful performance; across multiple tests, higher control performance should be expected from laboratories proficient with the test methods (Ingersoll etal. 2009). If a laboratory’s "typical” performance is not substantially above minimum acceptance criteria, it may be indicative of underlying problems with laboratory facilities, practices, or staff experience. Rather than trying to develop specifications for such things that may be difficult to quantify and evaluate, it makes sense to take a "performance-based” approach as described in USEPA (2000) and in ASTM (2011); that is, to set targets for the range of control performance observed over time, with the expectation that if a laboratory can maintain that level of performance over time, it is likely that the laboratory is functioning at a level appropriate to producing reliable results. It is important to note that these laboratory performance criteria are not intended to be used to judge the validity of individual tests; instead, these criteria represent what I believe should be a demonstration of overall laboratory capability. Acceptability of individual tests would still be judged against minimum standards included in the sediment toxicity test methods (or future modifications thereof). Table 1 below provides a summary of control performance characteristics I believe capable laboratories should be expected to achieve over the course of multiple tests. These characteristics are generally provided in two ways, relating to "minimum” and "average”performance. The "minimum” descriptors reflect performance or conditions that laboratories should only rarely fail to meet, while the "average” descriptors reflect performance that should be typical. It is clear that experience with a test method is essential for a laboratory to perform a sediment toxicity test procedure well. As with virtually any test method, control performance in sediment tests improves with experience as the individuals performing the test gain experience with the associated procedures. While all labs have to start somewhere, I think it is important that a laboratory have established a track record of test performance prior to undertaking an important study. This can be as simple as conducting "practice” or "control” tests which involve onlycontrol media. For the purposes of evaluating lab performance, I am assuming a laboratory that is actively conducting sediment toxicity tests will have completed at least 9 whole sediment toxicity tests(control plus test sediments) using the same species, test conditions, exposure duration, and endpoints during the past 3 years. The test data used for this evaluation should include, at aminimum, the 9 tests most recently conducted by the laboratory. If additional data prior to the 9 most recent tests are available, they may also be included. As a general rule, if laboratories havel ess experience with a test method than the 9 tests within the last three years, I would recommend they conduct additional control tests.

(USEPA 2000, ASTM 2010)




10-d Chironomus survival (%)
Not more than 15% of tests with
control survival <80%

Average ≥ 85%
10-d Chironomus startingending weight (mg
Not more than 15% of tests withending weight
<0.6 mg AFDW/individual
Average ≥ 0.8 mg
10-d Chironomus pupation
and/or emergence

Not more than 3% of total
organisms across all tests

Increased Minimum Acceptable Control Survival for 10-d Tests with Chironomus dilutus

Evaluation of performance data from multiple laboratories as shown that the 70% minimum control survival for the 10-d midge tests described by USEPA (2000) and ASTM (2011) is unnecessarily and inappropriately low (Ingersoll et al. 2009). I recommend that the minimum acceptable control survival for this test be increased to 80%.

Measurement (and reporting) of Starting Size of Test Organisms

While the existing test methods specify that subsamples of organisms used to initiate sediment toxicity tests be taken and measured to document starting size, it has been our experience that this practice is not always adhered to, or at least not reported if the measurements are made. I believe this is a very important measure for several reasons. For one, it provides documentation that the organisms were of an appropriate age/size at the beginning of the exposure. For midge starting size is important because if the organisms are too old/large at the beginning of theexposure, pupation may occur before the end of the exposure (see requirements for maximumstarting size for midge below). Furthermore, we believe that midge growth slows as the organisms approach pupation, such that differences in growth rates might be masked if the control organisms are in a period of little or no growth at the end of the exposure. In addition, knowing initial weight is important to assessing actual growth rather than just initial weight.

Maximum Starting Size for Midge in 10-d Tests

Existing guidance for the 10-d midge test specifies starting size for the organism in terms of age and/or instar. Recent experience suggests that 1) not many laboratories are actually determining instar distribution in starting organisms; and 2) that midges from different cultures reach different developmental stages at different times. For example, in our laboratory, a test starting with a 10-day old organism will likely show pupation by the end of a 10-d test, whereas in another laboratory there may be no pupation in the same time period. It is important to the proper execution of the test that it end before pupation. To that end, I believe a better approach to assuring an appropriate stage of development in midge larvae used to start the 10-d test is to make sure average starting ash-free dry weight (AFDW) does not exceed a maximum value; at this point, I think a mean starting weight of 0.12 mg AFDW/organism is a reasonable guideline, though considerably lower weights than this should be the rule. Regardless of starting sizes, laboratories observing pupation anything but rarely in 10-d tests must take steps to reduce the starting size/age/development of their midge.

Inclusion and Evaluation of a Quartz Sand Control

As discussed above, a goal for sediment toxicity testing is to have water and diet that is by itself sufficient to support test organisms. If organisms grow substantially better in field sediments than in a substrate like quartz sand, then it suggests that the organisms are relying on sedimentsto provide for a meaningful portion of their needs. This in turn means that differences in performance of Hyalella or midge among sediments could be confounded by what I’ll loosely call "nutritional quality” of the sediment, separate from the effects of sediment contaminants. In general, I believe control sediments should be sediments that readily support test organisms, but are not so rich that organisms grow better than might be typical of most field sediments. To evaluate this, I suggest that a laboratory have, at some point, conducted a study that includes botht heir regular control sediment and a quartz sand control. If growth in the control sediment exceeds that in the quartz sand control by more than 20%, I recommend that the laboratory consider an alternate control sediment. This is to avoid an unrealistic expectation of performance that might, among other things, cause uncontaminated but less "nutritious” sediments to appear toxic.

Transfer of Larval Midges to Test Chambers

Current methods do not provide much detail on the process of transferring midge larvae from culture media into the test chambers. Both our laboratory and the USGS-Columbia laboratory have found that control survival of midge in the 10-d midge test is improved if the midge larvae used for toxicity testing are selected from those remaining within their cases, rather than those that are found free of their case. In addition, these midges should be transferred to the test chambers by transferring the entire case (with the larva within) to the test chamber. While the exact reasons are not clear, it appears that midges found free of their cases are more likely to be stressed and show reduced survival in the sediment toxicity test. The USGS-Columbia laboratory in particular noted an immediate and marked increase in control survival rates from 70-80% to more than 90% when the "whole case transfer” procedure was adopted. Casual observations indicate that midges transferred in cases of culture sand readily move into the surrounding sediment, so there is little concern that the transfer of the sand case from the culture represents a meaningful "refuge” from exposure to the test sediment.