以多變量分析探討現行發布之生物毒性試驗

Abstract

水生生物毒性評估會因生物敏感性之差異而有低估毒性風險的發生，因此 不能只用單一種生物試驗來評估毒性。"組生物"試驗方法中包含多種生物試驗 方法，能得到較完整的毒性資訊。然而，我國環境檢驗所發布之環境生物毒性 檢驗方法中，並未考慮到物種敏感性及食物鏈之概念。因此，本研究目的除了 利用實驗及統計結果了解現行發布之生物毒性試驗方法是否足以有效評估汙 染物的危害性外，並為"組生物"試驗法選擇適當的生物測試方法，作為日後評 估環境毒物之參考依據。本實驗主要利用多變量統計法進行研究分析，為避免 統計結果缺乏準確性，研究以小樣本(二十種)及大樣本數(九十種)之有機物兩個 部份進行，所用到之生物試驗法包含密閉式與批次式月芽藻、水蚤、鯉魚、 Fathead minnow及海洋螢光菌六種試驗法，並將兩部份對應生物試驗所得之毒 性數據分別進行主成分分析與集群分析，以比較大小樣本數在統計上之差異性。 多變量分析結果顯示，密閉式月芽藻為本驗究中最具敏感性的試驗方法，且在 主成分分析中與水蚤皆為重要的成分變數；海洋螢光菌、鯉魚與Fathead minnow 之間具非常高之相關性，但會因有機物之特殊作用機制使其相關性改變。由於 大小樣本數之有機物統計結果相符合，依據主成分分析、群集分析及Pearson 相關係數分析之結果選擇的"組生物"包含：密閉式月芽藻、Daphnia magna及海 洋螢光菌或添加鯉魚試驗法，且根據最終之主成分分析結果顯示，以此"組生物 "來評估毒性並不會造成任何毒性評估資訊之損失。因此，密閉式月芽藻、 Daphnia magna及海洋螢光菌是"組生物"試驗法中合適的試驗生物。

In the aquatic toxicity testing, the potential risk of toxicants can be underestimated due to differences in species sensitivity; therefore, it can not only use single-species to assess the toxicity. Test batteries consist of several individual assays which have become more common and receive more complete toxicity information. However, from the testing methods for environmental biological toxicity of the Environmental Protection Administration of Taiwan, it does not consider the concept of species sensitivity and the different trophic levels. The aims of this study not only use experimental and statistical data to investigate whether the existing testing methods for environmental biological toxicity can assess the harmfulness of the pollution effectively, but also select an optimal battery of bioassays as a source references to evaluate ecotoxicity in the future. This study utilize mainly the multivariate analyses, to avoid the insufficient accuracy of the statistical result, the experiment is proceeding from two parts: small sample (Twenty chemicals) and large sample (ninety chemicals). The six bioassays include closed-system and open-system algal(Pseudokirchneriella subcapitata), Daphnia magna, Cyprinus carpio, Fathead minnow and Microtox, then using Principal Component Analysis (PCA) and Cluster Analysis(CA) separately from the toxicity data of the two corresponding biological test. These multivariate analysis clearly showed that closed-system algal is the most sensitive testing method in this study, and is the same as Daphnia magna in the PCA that both are the important component variable. The correlation among Microtox, Cyprinus carpio and Fathead minnow is high, but it changes according to the special mode of action. PCA, CA and Pearson allowed for the selection of the following test battery：closed-system algal、Daphnia magna and Microtox or in addition Cyprinus carpio test. And the final PCA shows that the number of bioassay can be reduced without significantly losing information. Thus, closed-system algal, Daphnia magna and Microtox are the ideal bioassays in the test battery.