A comprehensive critical review was undertaken aiming to compare the intrinsic sensitivity of terrestrial plant species (crop species and noncrop wild species) with published literature and unpublished proprietary data generated for the registration of plant protection products (PPPs), and a database was compiled. Data were assessed to answer the question whether crops differ from noncrop plants in their intrinsic sensitivity to PPPs. Endpoints were assessed considering further potentially relevant parameters by means of different methods, including a quotient approach, in which overall crop endpoints were divided by matching wild species endpoints. Quotients above 1 indicated that wild species were more sensitive than crops, quotients below 1 the opposite. Further methods included a multiple regression analysis and different approaches to assess the statistical power. The overall finding was that there were no consistent differences in sensitivity between wild plant species and crop species, based on ER50, ER25, and ER10 vegetative endpoints (the largest fraction of data). This was also true when censored endpoints, seedling emergence data, and other measured variables such as shoot height were included. Statistically significant differences occurred in both directions and were balanced, that is, there was no clear trend for either crops or noncrop species to be more sensitive than the other. On the basis of multivariate regression analysis, crops were found to be significantly more sensitive than wild plant species, albeit by a small margin (factor %1.4). Minimum detectable difference (MDD) analysis and multivariate regression analysis of modified datasets indicated that when using a data set of this size and heterogeneity, any dissimilarity between crop and wild species was detectable if exceeding a factor of 1.4 in either direction. For the taxonomic groups assessed here (i.e., with data), no intrinsic difference in sensitivity to PPPs between crop species and wild plant species was found. Integr Environ Assess Manag 2019;15:176-189. C 2018 SETAC

To investigate whether vegetative endpoints are protective of reproductive endpoints in terrestrial plant risk assessments (RAs) for authorization of plant protection products (PPPs), we assessed differences in sensitivity to herbicides between these parameters. Published literature and unpublished proprietary data generated for the registration of PPPs were used to compile a database. If reproductive endpoints were systematically more sensitive than the vegetative endpoints on which regulatory decisions are presently based, a concern could be raised about the protectiveness of the current RA process. Vegetative and reproductive endpoints were assessed considering further potentially relevant parameters. Reproductive endpoints were compared with vegetative endpoints of juvenile plants or with those of mature plants. Direct comparison by substance–species combination proved to be most adequate and was used to calculate quotients by effect level. In addition, we assessed the spread between different effect levels, estimating by which factor the conservatism would increase if effect rate (ER)50 endpoints were replaced by ER25 or ER10 endpoints with otherwise unchanged test parameters. Reproductive endpoints were found to be similarly sensitive as vegetative endpoints derived in nontarget terrestrial plant (NTTP) studies conducted following Organisation for Economic Co‐operation and Development (OECD) or US Office of Chemical Safety and Pollution Prevention (OCSPP) guidelines. A switch from vegetative to reproductive endpoints would therefore not significantly change the conservatism (less than a factor of 1.5), whereas the change from ER50 to ER10 would (by a factor of ~5 to 6). However, because ecotoxicological tests on terrestrial plants bear intrinsic high variability that prevents the reliable detection of effects at the 10% effect level, ER10 endpoints are not a reliable basis for RA. No particular family, genus, or species with clusters of distinctly insensitive vegetative and sensitive reproductive endpoints could be identified that would call for regular testing of reproductive endpoints. Also, from the data set available, no specific herbicidal modes of action could be singled out for acting particularly on reproductive endpoints. Integr Environ Assess Manag 2020;16:166–183. © 2019 SETAC