# SSD new batch (operator Taildrop 2026-06-14) — close-read excerpts

*Seven papers retrieved 2026-06-14, title-verified, filed in `../papers/`, and close-read. Verbatim quotes below
(grade **[P]**, read from the local PDFs). These upgraded the essay [*The Fifth Percentile*] critique section
from abstract-level **[2]** to held primaries, and added the Dutch-rationale and self-validation strands.*

## Forbes & Calow (2002) — "SSDs Revisited: A Critical Appraisal," *HERA* 8(3):473–492
The flagship critique. Abstract + intro:
> "if we could determine a community sensitivity distribution, this would provide a better estimate of an
> ecologically relevant effects threshold… However, the distributions generated are typically based on
> **haphazard collections of species and endpoints** and by adjusting these to reflect more realistic trophic
> structures we show that effects thresholds can be shifted **but in a direction and to an extent that is not
> predictable**. Despite claims that the SSD approach uses all available data… in certain frequently used
> applications **only a small fraction of the species going into the SSD determine the effects threshold**."

> "the SSDs that are often generated **do not represent any known community, but are often interpreted as if they
> do**." *(the sharpest line; used in essay §V.)*

Assumption they flag as false: "T1. Interactions between species do not influence the sensitivity distribution"
(knock-on / trophic / competitive effects). Conclusion echoes Forbes & Forbes 1993 — the real uncertainties are
"obscured by the apparent sophistication of the models and their outputs."

## Forbes & Calow (2003) — "Does Ecotoxicology Inform Ecological Risk Assessment?" *Environ. Sci. Technol.* 37:1544 (Viewpoint)
Risk assessments "usually focus on protecting species composition because that should ensure the protection of
ecosystems" — the inference (species list → working ecosystem) the authors question. Frames the field as possibly
"too simplistic to do the job effectively."

## Newman et al. (2000) — *Environ. Toxicol. Chem.* 19(2):508–515
> "The increasingly common assumption that a lognormal model best fits these data was not supported. **Fifteen
> data sets failed a formal test of conformity to a lognormal distribution**… An alternate **bootstrap** method
> provided accurate estimates… [required sample sizes] ranged from **15 to 55, and had a median of 30**
> species-sensitivity values. These sample sizes are **higher than those suggested in recent regulatory
> documents**."

## Maltby et al. (2005) — *Environ. Toxicol. Chem.* 24(2):379–388
> "The **taxonomic composition** of the species assemblage used to construct the SSD does have a significant
> influence on the assessment of hazard, but the **habitat and geographical distribution of the species do not**.
> Moreover, SSDs constructed using species recommended in test guidelines **did not differ significantly** from
> those constructed using nonrecommended species."

Recommends "a **safety factor of at least five** should be applied to the median HC5" for continuous/repeated
exposure. *(Note: a precise selection-bias result — which TAXA matter, not where they live.)*

## Van Leeuwen (1990) — *Environ. Management* 14:779–792 (Dutch DG Environmental Protection)
The Dutch government's contemporaneous rationale; references **"Premises for Risk Management"** and the **National
Environmental Policy Plan of the Netherlands (1989)**, with protection extended "to **95% of the species in
ecosystems**." Critically reviews the Health Council's extrapolation methods and Van Straalen & Denneman (1989).
*(Upgrades the essay's §II "Premises adopted the 5%" claim from [synthesis] to a held government primary.)*

## Emans et al. (RIVM, 1993) — "Validation of Some Extrapolation Methods Used for Effect Assessment," *ETC* 12:2139–2154
RIVM's own validation (the Dutch counterpart to P&G's Versteeg 1999):
> "a validation of extrapolation methods was carried out by comparing NOECs derived from multiple-species
> (semi-) field experiments with extrapolated values… it is concluded that **single-species toxicity data can be
> used to derive "safe" values**… [Aldenberg & Slob and Wagner &] Lokke, both with a **95% protection level and a
> 50% confidence level**."

Restricted by an admitted "paucity of data"; "safe" in the authors' own quotation marks. *(The method's own
authors validating the method — used in essay §VII.)*

## Posthuma, Suter & Traas (eds.) (2002) — *Species Sensitivity Distributions in Ecotoxicology* (SETAC/Lewis, 617 pp)
The canonical reference. **Ch. 2, Suter, "North American History of SSDs"** (pp. 11–17): "EPA staff members
decided in **1978** to replace the use of [application factors]…"; the method for "5th percentiles of SSDs (HC5)
was **repeatedly revised** until the U.S. EPA (1985) [Guidelines]" set the Final Acute Value. **Ch. 3** is the
"European History of SSDs." *(Insider history; dates EPA's percentile turn — essay §II.)*