For more than 25 years, there have been reports that effluents from pulp and paper mills affect fish reproduction. Studies involving wild fish, in situ experiments, and laboratory in vivo tests conducted in Sweden, Canada, Finland, the United States, and New Zealand have documented reductions in sex steroid hormone levels, gonad size and fecundity, alterations in secondary sex characteristics, and delayed sexual maturity associated with exposure to mill effluents.
Since the late 1980s, the global industry has made significant changes in mill operating conditions; these changes were designed, in part, to mitigate environmental impacts and to comply with regulatory requirements. The elimination of polychlorinated dioxins and furans from mill effluents through the reduced use of molecular chlorine in bleach plants, the installation of biotreatment systems, and the implementation of color-removal strategies have combined to produce effluents with significantly reduced loadings of organochlorines, biological oxygen demand (BOD), acute lethal toxicity, suspended solids, and color [1–4].
Despite these changes, effects on fish reproduction continue to be reported in Canada , the United States , Sweden [7,8] Finland , and New Zealand , with effects also noted recently in countries with emerging pulp and paper industries, such as Chile [11,12]. Perhaps the most up-to-date trends are evident from the regulatory Environmental Effects Monitoring (EEM) program implemented in Canada during the mid-1990s. As part of the EEM program, mills are required to monitor fish and benthos in their receiving environments in three-year cycles. This information is then used to assess the adequacy of the effluent regulations in protecting the environment on a site-specific basis . Within the EEM, an effect on the fish population and on the benthic invertebrate community is defined as a statistically significant difference in a measured parameter between an area exposed to effluent and a reference area or as a statistically significant gradient within the exposure area . If effects are observed above a critical effect size, mills are required to undertake an investigation of cause  and, pending regulatory approval, an investigation of solutions to remedy the situation.
Information from the first three cycles of the EEM in Canada has revealed two national average response patterns in fishes—namely, enhanced growth and enhanced condition that can be attributed to nutrient enrichment and metabolic disruption . Metabolic disruption is characterized by increased body condition, increased liver size, and decreased gonad size (all relative to reference), and metabolic disruption has been documented in mills employing a range of pulp production and effluent treatment types discharging into a variety of environments (e.g., riverine, lake, and estuarine). The consistency of this response pattern as well as the evidence from other studies worldwide has highlighted the necessity for mitigation. Because the sources, identities, and biological mechanisms associated with fish reproductive impairment, however, remain unresolved, little information is available about what mills can do to improve effluent quality in relation to fish reproduction.
Table Table 1.. Summary of effects of pulp mill effluents on wild fish
|Florida (USA) rivers, late 1970s to present||Bleached kraft||˜80||Mosquito fish and other Poicillidae||Masculinization of females based on secondary sex characteristics||[5,57,117,121]|
|Baltic Sea (Sweden), 1980s to 1990s||Bleached and unbleached kraft||<1–5||Mainly perch (Perca fluviatilis) and roach (Rutilus rutilus)||Reduced gonad size; larval mortality||[16,20,135,136]|
|Canadian freshwaters (>15 sites), late 1980s to present||Bleached kraft, mechanical/sulfite, sulfite TMP, BCTMP||0.2–22 (overall average. 1)||>10 species, mainly white sucker (Catostomus commersoni)||Reduced gonad size, circulating sex hormones, and fecundity; delayed maturity; changes in secondary sex characteristics|||
|Baltic Sea (Sweden), late 1990s to 2002||Bleached kraft||<1–5||Eelpout (Zoarces vivparous)||Greater proportion of male embryos|||
|North Carolina/Tennessee (USA) rivers, 1989 and 1990||Bleached kraft||˜75||Redbreast sunfish (Lepomis auritus)||Lower serum levels of estradiol and increased incidence of atretic vitellogenic oocytes in females|||
|Florida (USA) rivers, late 1990s||Bleached kraft||40–80||Largemouth bass (Micropterus salmoides floridanus)||Reduced gonad size, lower plasma sex hormones, and reduced vitellogenin in females|||
|Waikato River (New Zealand), 2002||Bleached kraft||˜50||Brown bullhead (Ameiurus nebulosis)||Lower serum levels of steroid hormones, no change in gonad size|||
|Lake Saimaa (Finland), 1995 and 1996||Bleached kraft||1–4||European perch (Perca fluviatilis L.) and roach (Rutilus rutilus L.)||Decreased gonad size and plasma sex steroid hormones in perch only||[9,139]|
The purpose of this review is to provide an overview of the primary reproductive effects observed in fish exposed to pulp and paper mill effluents, the current state of knowledge regarding the sources/causes of reproductive effects, and the consequences of changes in mill operating conditions in relation to fish reproductive impacts. It is hoped that this information will facilitate research leading to the formulation of cost-effective strategies for eliminating reproductive effects associated with the effluent discharges, thereby ensuring the sustainability of both fish populations and the pulp and paper industry.
REPRODUCTIVE RESPONSES IN FISH EXPOSED TO PULP MILL EFFLUENTS
The first indications linking mill effluents to reproductive effects were noted in wild fish (summarized in Table 1). Subsequently, fish exposures in laboratory, mesocosm, and in situ studies also demonstrated that mill effluents can influence a variety of reproductive endpoints (Table 2). Whereas some results vary by species and from study to study, effluents can affect fish reproduction in multiple ways, with the most notable and consistent responses including decreased gonad size, decreased production and/or levels of gonadal sex steroids, hormone–receptor interactions, altered expression of secondary sex characteristics, and decreased egg production.
Decreased gonad size
Numerous studies have shown that gonad size is decreased in wild fish exposed to pulp and paper mill effluents (Table 1). It is one of the most consistent responses observed, and it has been reported in Sweden , Finland , Canada , and New Zealand . Decreased gonad size has coincided with delayed sexual maturation offish in Canada [18,19] and Sweden .
In Canada, reductions in gonad size have been consistently measured in white sucker (Catostomus commersoni) exposed to effluent from a bleached kraft mill located at Jackfish Bay (ON, Canada) since the late 1980s . Decreased gonad size also has been consistently found in perch (Perca fluviatilis) collected near the bleached kraft mill located at Norrsundet (Sweden) . Reductions of gonad size have been observed in fish at other Canadian mills as well [22,23]. Furthermore, three successive cycles of Canadian EEM studies from 1992 to 2003 evaluated multiple fish species and determined an average decrease in gonad size of both sexes of fish at approximately 60 mills . Decreased gonad size, combined with increased energy use and storage, had been interpreted as a form of “metabolic disruption” . Although observed primarily in wild fish, gonad size reductions have been observed in laboratory-based, life-cycle studies with fathead minnow (Pimephales promelas)  and in mesocosm (i.e., artificial stream) studies with mummichog (Fundulus heteroclitus) .
Decreased production of gonadal sex steroids
In many studies, decreases in circulating levels of gonadal sex steroids (testosterone, 17β-estradiol, and 11-ketotestosterone) coincide with decreased gonad size. Studies with white sucker exposed to bleached kraft mill effluent (BKME) at Jack-fish Bay showed both sexes had lower levels of sex hormones in effluent-exposed zones , and a number of sites within the pituitary–gonadal axis were affected. This included reduced levels of gonadotropin II (i.e., luteinizing hormone), reduced ovarian steroid biosynthetic capacity, and altered peripheral steroid metabolism . Interestingly, ovarian prostaglandin production was not affected in these fish, suggesting that the effects were specific to steroid biosynthesis and not a general response of the ovary to effluent exposure. More recent tests with Canadian bleached sulfite mill effluent (BSME) showed inhibition of steroidogenesis, as evidenced by decreases in pregnenolone concentrations in female rainbow trout (Oncorhynchus mykiss) . Decreases in circulating and gonadal production of testosterone also have been observed in the estuarine mummichog after exposure to primary-treated BKME  and in-mill process streams .
Table Table 2.. Summary of effects of pulp mill effluents from in vivo laboratory, in situ, and mesocosm fish studies
|Fathead minnow (Pimphelas promelas)||BKME, BSME, and TMP (Canada and United States)||Life cycle (egg to sexual maturity); 6 months||Egg production at 1.7–80%; also changes in secondary sex characteristics, including masculinization and feminization, changes in sex hormones, delayed sexual maturity||[61–66,109–113]|
| ||BKME, TMP, and MP (Canada)||Adult reproduction; 21 d||Egg production and VTG induction at 20%||[1,54]|
|Mosquito fish (Gambusia affinis)||BKME and TMP (New Zealand)||Adult exposure; 21 d||Masculinization at 100%|||
|Shortfin eel (Anguilla australis)||BKME, TMP, and CTMP (New Zealand)||Juvenile in situ exposure; 21 d||Increased plasma estradiol and testosterone at ˜10%|||
|Guppy (Poecilia promelas)||BKME (Sweden)||Adult exposure; 42 d||Masculinization at 5–25%|||
|Mummichog (Fundulus heteroclitus)||BKME (Canada)||Adult exposure; 7–57 d||Reduced plasma testosterone at 1 and 5%||[30,31]|
|Goldfish (Carassius auratus)||BSME, BKME, and TMP (Canada)||Adult exposure; 8, 16, and 21 d||Reduced circulating sex hormones and gonadal hormone production at 25–100%||[96,123]|
|Rainbow trout (Oncorhynchus mykiss)||BKME and TMP (New Zealand)||Maturing (two years or older); 8 months||Reduced gonads, testosterone, and estradiol in females at 12%; reduced larval size in progeny of exposed adults.||[48,124]|
|Largemouth bass (Micropterus salmoides)||BKME and unbleached KME (USA)||Adult (1.5 years); 28– 56 d||Reduced sex hormones and gonad size at ≥ 20%||[50,141]|
|Three-spined stickleback (Gasterosteus aculeatus)||Primary BKME (Sweden)||Adult females; 42 d||Masculinization (increased spiggin and kidney epithelial cell height) at 10%|||
Other work has focused on potential mechanisms underlying the altered steroid biosynthetic capacities of effluent-exposed fish. Studies with white sucker identified a number of sites within the steroid biosynthetic pathway effected by effluent exposure . By measuring the synthesis of the major steroid intermediates in the biosynthetic pathway, it was possible to identify sites where the activities of specific enzymes were altered. Unfortunately, these sites changed with the reproductive state of the fish and did not account for all the reductions. It was more probable that reductions in the availability of the primary substrate cholesterol could account for the major reductions in the production of steroids in exposed fish . Ovarian tissues from these fish also have reduced mRNA expression of the steroid acute regulatory protein (StAR) that is responsible for mobilization of cholesterol across the mitochondrial membrane (M. McMaster, Environment Canada, Burlington, ON, unpublished data).
Other studies have shown that sitosterol, the dominant plant sterol found in mill effluents, affects steroid biosynthesis in gonadal tissues by interfering with cholesterol mobilization and its conversion to pregnenolone . In the gonad, the StAR protein normally facilitates cholesterol mobilization to the mitochondria, where it is converted into pregnenolone by the cytochrome P450 cholesterol side-chain cleavage enzyme (P450scc). Emerging evidence suggests that expression of StAR and P450scc is reduced in fish exposed to aryl hydrocarbon–receptor ligands , which are present in mill effluents. Another potential pathway may involve the reduced levels of luteinizing hormone observed in white sucker exposed to BKME  and the finding that StAR protein levels are regulated by gonadotropins in zebrafish [28,35]. Recently, StAR mRNA transcript levels and plasma testosterone levels have been shown to be reduced in male goldfish (Carassius auratus) after five months of exposure to sitosterol .
Pulp mill effluents contain ligands for nuclear sex steroid receptors and the plasma sex steroid–binding protein, thereby having the potential to affect steroid hormone signaling and transport in fish. Ligands for estrogen and androgen receptors and the sex steroid–binding protein have been detected in hepatic tissues of fish exposed to effluent for 4 d at the Jackfish Bay mill , at a Canadian bleached sulfite mill , and in wild fish collected during the spring spawning migration below an additional Canadian bleached kraft mill . Effluents from each of these mills have been shown to cause reproductive alterations in wild fish. Ligands for the androgen receptors have been detected in Swedish kraft mill effluent  that causes male-biased sex ratios of eelpout (Zoarces viviparous)  and have been linked to masculinization of mosquito fish (Gambusia affinis) in Florida (USA) [6,41] and New Zealand .
Table Table 3.. Summary of studies investigating the effects of pulp mill effluents on vitellogenin (VTG) in fish
|BKME||White sucker||Field (Canada)||Life||M||↓||[44,142]|
| || || || ||F||0|| |
|MP||White sucker||Field (Canada)||Life||M||0|||
| || || || ||F||0|| |
|BKME||Largemouth bass||Field (USA)||Life||M||0|||
| || || || ||F||0|| |
|BKME||Largemouth bass||Mesocosm (USA)||28–56 d||M||0|||
| || || || ||F||↓|| |
|BKME||Bluegill||Mesocosm (USA)||8 months||M and F||0|||
| || || || ||F||↓|| |
| || || || ||F||0|| |
| || || || ||F||0|| |
| || || || ||F||0|| |
|BKME||Whitefish||Field (Finland)||30 d||I||↑|||
|BKME||Whitefish||Field (Finland)||30 d||I||0|||
|BKME||Whitefish||Field; Finland||30 d||I||0|||
|TMP and kraft||Rainbow trout||Caged (Sweden)||21 d||I||0|||
|BSME||Rainbow trout||Laboratory (Canada)||21 d||I||↑|||
|BKME||Rainbow trout||Laboratory (Canada)||21 d||I||↑|||
|BSME||Rainbow trout||Laboratory (Canada)||21 d||I||↑|||
|BKME||Rainbow trout||Laboratory (Canada)||21 d||I||0|||
|BKME and TMP||Rainbow trout||Laboratory (New Zealand)||60 d||M||0|||
|BKME and TMP||Rainbow trout||Laboratory (New Zealand)||60 d||M||↑|||
| || || || || ||F||↓|
|BKME and TMP||Rainbow trout||Mesocosm and laboratory (New Zealand)||7, 14, 21, 28, and 320 d||I||0|||
|BKME||Brown trout||Laboratory (Canada)||15 d||I||0|||
|BKME||Brown trout||Laboratory (Canada)||28 d||I||0|||
|TMP||Fathead minnow||Laboratory (Canada)||21 d||I||↑|||
|BKME||Fathead minnow||Laboratory (Canada)||21 d||I||↑|||
|MP||Fathead minnow||Laboratory (Canada)||21 d||I||↑|||
|TMP||Fathead minnow||Laboratory (Canada)||21 d||I||0|||
|BKME||Fathead minnow||Laboratory (Canada)||21 d||I||0|||
|MP||Fathead minnow||Laboratory (Canada)||21 d||I||0|||
|TMP||Fathead minnow||Laboratory (Canada)||21 d||I||↑|||
|BKME||Fathead minnow||Laboratory (Canada)||21 d||I||↑|||
|MP||Fathead minnow||Laboratory (Canada)||21 d||I||↑|||
|TMP||Fathead minnow||Laboratory (Canada)||21 d||I||0|||
|BKME||Fathead minnow||Laboratory (Canada)||21 d||I||0|||
|BKME||Fathead minnow||Laboratory (Canada)||5 months||M and F||↑|||
|River sediment||Rainbow trout||Laboratory (Chile)||29 d||I||↑|||
|BKME||Rainbow trout||In situ (Chile)||11, 21, and 30 d||F||↑|||
|BKME, TMP, BCTMP, and BSME||Rainbow trout hepatocytes||Laboratory (Canada)||4 d (extract incubation)|| ||↑, 0,↓|||
Hormone–receptor interactions also are evident in the expression of estrogenic responses. Increased amounts of the yolk precursor vitellogenin (VTG) are detected in males following exposure to estrogen agonists as well as to androgenic substances that are metabolized to estrogens . In the case of pulp mill effluents, the available information concerning effluent effects on VTG induction is highly variable and may be related, in part, to exposure concentrations and differences in species sensitivities (Table 3). Several studies have found no effects of effluent exposure on VTG production, and other studies have shown that circulating levels of VTG in females is depressed . For example, in the Moose River basin of Canada, circulating levels of VTG were reduced in white sucker females downstream from a bleached kraft mill, corresponding to reduced circulating levels of 17β-estradiol [43,44]. No VTG induction was observed in immature brown trout (Salmo trutta) exposed to two Canadian kraft effluents, one at 25 to 100% for 15 d and the other at 100% for 28 d. . Elevated VTG was not seen in roach (Rutilus rutilus) caged downstream of two elemental chlorine-free (ECF) mills in Finland, and VTG in females was decreased in perch downstream from one of the two mills . Similarly, no VTG induction has been observed in sexually mature rainbow trout exposed to mill effluents in Sweden . In New Zealand, a BKME/thermo-mechanical (TMP) effluent was shown to induce VTG production in only one of four experiments [47-49], suggesting temporal fluctuations in effluent compositions. In the United States, largemouth bass (Micropterus salmoides; mesocosm and wild) exposed to BKME in Florida [50,51] and bluegill (Lepomis macrochirus) exposed in artificial streams for eight months to 12 and 30% BKME from a mill in North Carolina (USA)  also showed no evidence of estrogenic effects as measured by VTG induction.
Conversely, examples also exist in which the induction of VTG was associated with effluent exposure (Table 3). Sexually immature rainbow trout exposed in the laboratory to effluents from Canadian bleached sulfite and bleached kraft mills exhibited significant VTG induction . In tests with 11 Canadian mill effluents, male fathead minnows exposed to 20 or 40% effluent from TMP, kraft, and multiprocess (both chemical and mechanical pulping) mills showed VTG induction as the most frequent response [1,54]. Emerging information regarding the potential of effluents from South American mills to affect fish reproduction showed a twofold induction of plasma VTG in rainbow trout exposed for 29 d in the laboratory to sediments collected below the discharge of four Chilean pulp and paper mills in the Biobio River . The increased level of VTG also was associated with increased gonad size and the presence of more mature ovarian follicles in females. Immature female trout caged for 21 d at the same locations where the sediments were collected exhibited four- to fivefold increases in VTG levels that coincided with enhanced gonadal maturation (presence of vitellogenic oocytes) .
Other studies have shown that different pulp mill effluents may contain compounds that exhibit estrogenic or antiestrogenic activity. Biotreated effluent extracts from Canadian mills of several process types were shown to increase VTG production in primary cultures of rainbow trout hepatocytes, whereas others antagonized the actions of 17β-estradiol on VTG induction . These observations, in combination with exposure concentrations and species sensitivities, may account for the discrepancies noted in the literature.
Altered expression of secondary sex characteristics
The earliest evidence of pulp mill effluent-induced changes in secondary sex characteristics comes from the 1980s: Female mosquito fish in the Fenholloway River of Florida were found to be masculinized [56,57]. One of the most conspicuous indicators for masculinization of female mosquito fish is the formation of a gonopodium, which involves elongation of the anal fin that typifies sexual development in males. This has been observed in mosquito fish exposed in the laboratory to the kraft effluent discharged to the Fenholloway River [57,58] and, more recently, to ECF kraft/TMP effluent from New Zealand, where effluent filtration eliminated the masculinization effect in laboratory exposures . Follow-up work at the New Zealand mill showed no masculinization in wild mosquito fish and that effluent-associated suppression of in vitro production of ovarian sex steroids was not related to masculinization . A significant amount of work has been directed at the role of microbial transformation of phytosterols contributing to masculinization and is discussed below (see Experiments with individual compounds and Toxicity identification evaluation). An example of reduced expression of secondary sex characteristics are white sucker collected from Jackfish Bay that showed reduced numbers of tubercles in males exposed to primary-treated effluent . Interestingly, after installation of secondary treatment, tubercles were noted in some females, and some recovery, although not complete, was found in males .
Effects on secondary sex characteristics also have been noted in fathead minnow and represent one of the most sensitive responses of this species to effluent exposure in the laboratory . Responses that have been observed in laboratory studies include delayed development of sex characteristics, demasculinization of male fish, feminization of male fish (e.g., ovipositor development), and masculinization of female fish (tubercles and dorsal fin dot) [25,61–63].
Decreased egg production
Controlled laboratory studies have consistently found that mill effluents have the capacity to affect egg production. Life-cycle studies involving fathead minnows exposed to effluents from U.S. and Canadian bleached kraft mills [63–66], unbleached kraft mills [64,65], and one bleached sulfite mill  have found reductions in the number of eggs produced.
Effects on egg production commonly have been assessed with laboratory studies using fathead minnows in life-cycle assays (from eggs to reproductively mature adults; approximately five months) and in abbreviated life-cycle tests of 28 d. Abbreviated tests involve a pre-exposure period, from which spawning pairs of fish are selected, followed by an exposure phase . Egg production was negatively affected at concentrations of effluent lower than approximately half the effluent concentrations that negatively impacted in vitro steroid production .
Employing a 28-d adult reproductive test , egg production recently has been used to survey the effects of 11 Canadian effluents [1,54] as well as to track sources of endocrine-disrupting substances at the bleached kraft mill in Jackfish Bay [68,69]. In the 11-mill survey, in which effluent concentrations did not exceed 40%, egg production was halted completely by one of the effluents. For the Jackfish Bay mill, on-site exposures subsequently were used to test the effects of various process streams within the mill. This showed that both the combined mill effluent (before secondary treatment) and the combined alkaline stream caused decreased spawning events and decreased egg production.
SOURCE AND NATURE OF SUSPECTED CAUSATIVE AGENTS
Pulp and paper mill effluents are complex matrices containing material from wood (e.g., extractives, phytosterols, and trace metals), process derivatives or compounds formed during pulping/bleaching (e.g., dimethyl disulfide formed during kraft pulping), additives (e.g., polymeric formulations used as retention aids in papermaking), and if the effluent is biotreated, (partially) biodegraded products of the above. Much of the chemical characterization of mill effluents was performed during the 1980s and 1990s [70–73], with no reports of detailed chemical characterization of effluents appearing since the mid-1990s. Following process modifications over the last decade, however, effluent compositions have changed markedly since the previous characterizations, and this lack of information impedes efforts to establish cause-and-effect relationships with present-day effluents.
Insight regarding the identities and sources of chemicals affecting fish reproductive homeostasis has arisen indirectly from experiments with individual effluent extractives and long-term field studies. Several attempts also have been made to identify directly the source and nature of substances in mill effluents impacting fish reproduction, classified as toxicity source evaluation (TSE) and toxicity identification evaluation (TIE) studies. The TSE studies involve determinations of individual process wastewaters within the mill that are causing the effect (see Toxicity source evaluation), and the TIE studies describe effluent (or individual waste stream) manipulations that lead to the isolation, identification, and confirmation of the suspected causative agents (see Toxicity identification evaluation).
Experiments with individual compounds
Many studies have examined the reproductive effects of individual effluent extractives, and much of this work has been reviewed  (http://ec.europa.eu/environment/endocrine/documents/mrcreport.pdf). These extractives include resin acids (e.g., abietic acid), isoflavonoids (genistein), and phyto-sterols (sitosterol and stigmastanol). All originate from wood and may end up in the final discharge either unaltered or as a biotransformation product of microbial activity. Several researchers have hypothesized that because the chemical structures of these compounds are similar to those of endogenous steroids, these wood components may be causing the reproductive effects via interference with receptor-mediated pathways. It is important to recognize that although these studies have tested the effects of selected compounds on an individual basis, the effects of mixture interactions that could occur in actual exposures have not been considered.
The greatest body of work to date involves phytosterols, mainly sitosterol and, to a lesser extent, stigmastanol. Various formulations of sitosterol have been found to affect various aspects of fish reproduction (e.g., steroid production and VTG induction) in goldfish and rainbow trout [53,75-77]. Detailed follow-up studies have found that the actions of sitosterol leading to depressions in hormone biosynthesis are mediated through effects within the steroid biosynthetic pathway [33,78]. When sitosterol exposures were performed at typical North American mill effluent concentrations  in parallel with effluent from a Canadian bleached kraft and a bleached sulfite mill, however, plasma cholesterol was reduced by sitosterol but not by the mill effluents , suggesting that sitosterol was not the direct source of the effects.
Several long-term fish exposures with sitosterol have been conducted. A phytosterol preparation from wood, containing mainly sitosterol, was used in a multigenerational test with zebrafish (Danio rerio) . Exposure to sitosterol at 10 μg/L in the mixture did not affect spawning success. Induction of VTG was noted, however, and the sex ratio of the F1 and F2 generations was altered, with evidence for masculinization of the former and feminization of the latter. This suggested that the phytosterol preparation had both estrogenic and androgenic effects. In another experiment, maturing brown trout (Salmo trutta lacustris) were exposed to 10 and 20 μg/L solutions of phytosterol (mainly sitosterol) . After artificial fertilization, the development of the next generation was followed. Exposure to phytosterols affected the next generation, as evidenced by higher prevalence of disease, deformities, and mortality.
Work on understanding the causes of masculinization associated with sterol exposure dates back to the mid-1980s, with conflicting reports concerning the role of microbial transformation in the effects. During an initial study attempting to explain the cause of masculinization, mosquito fish were exposed in the laboratory to predominantly sitosterol or stigmastanol with associated sterol impurities . Fish also were exposed to the phytosterols mixed with Mycobacterium smegmatis as well as to the bacteria alone. No evidence was found of females being masculinized when exposed to the bacteria alone or to the phytosterols alone. Conversely, masculinization was observed in females exposed to the phytosterol/bacterial mixtures, with slightly greater potency observed for the stig-mastanol/bacteria mixture . This indicated that the active compounds are not the phytosterols themselves but, rather, their bacterial degradation products. The actual degradation products were not identified, which has led to much speculation regarding their identities. Other studies with stigmastanol using fathead minnow life-cycle exposures have shown that concentrations an order of magnitude higher than those typically found in final effluents (˜73 μg/L) caused no effect, and no correlations have been found between threshold effects on egg production and other phytosterols, including sitosterol, campesterol, stigmasterol, and total phytosterols . Recent work from New Zealand demonstrates that sterols in effluents and their oxidation products resulting from chlorine dioxide bleaching are not androgenic . Other studies have determined that multiple compounds are functioning as ligands for the androgen receptor and that these likely are involved in the responses (see Toxicity identification evaluation).
Resin acids have been tested for their ability to affect fish reproduction because of their acute toxicity and ubiquitous nature in effluents from softwood mills. Abietic acid was reported to have weak estrogenic activity [84,85], although the purity of the product tested was low. Because resin acids typically are reduced by more than 90% in biotreatment systems , their probable role in causing reproductive effects with fish exposure to biotreated effluent generally is regarded as minor. Genistein was found at concentrations of approximately 13 and 11 μg/L at a Canadian kraft mill before and after biotreatment, respectively . In tests with the Japanese medaka (Oryzias latipes), genistein as well as equol, another isoflavonoid, were found to cause feminization of secondary sex characteristics in the males as well as a variety of ovarian alterations in the females . Evidence of intersex (testisova) also was found in males. These effects, however, occurred at genistein concentrations (e.g., 1 mg/L) 100-fold higher than those found in the effluent.
Pulp mill effluents also may contain additives; however, to date, only nonylphenol has been examined for its potential to affect fish reproduction. Nonylphenol is a biodegradation product of alkyl phenol ethoxylates present in formulations used for pitch dispersants, felt washers, cleaners, defoamers, and deinking agents, and it has been shown to possess estrogenic properties . The pulp and paper industry did use alkyl-phenol ethoxylate products, but the use of such products was voluntarily discontinued . As such, nonylphenols likely are not the cause of the effects in the effluents discharged by present-day mills.
Indirect evidence suggesting cause
Based on field studies that monitored fish as mills implemented process changes (e.g., before and after the installation of effluent biotreatment; see Role of effluent biotreatment), several inferences were drawn about the nature of the effluent components associated with changes in circulating steroid hormones . These include the following: Biotreatment does not completely eliminate the causative agents, the causative substances affect fish through waterborne exposure, and continuous effluent exposure was necessary to maintain the effects. Finally, effects on steroid levels are not long lasting—that is, once removed from effluent, fish appear to recover, and the effects caused by discharges from modernized or current mills are unlikely to be caused by highly substituted, chlorine-containing compounds.
Attempts to correlate water chemistry with biological effects have been largely unsuccessful and underscore the complexity of establishing cause and effect. For example, water samples from sites on three rivers in Florida receiving effluents from three bleached kraft mills were analyzed for chlorinated phenolics, resin/fatty acids, phytosterols, and various group parameters, such as polyphenolics and condensable tannins . The masculinization response in female mosquito fish was greatest at sites where resin/fatty acids and sitosterol peaked in concentration; however, the opposite was found at another site, suggesting that these compounds are not involved with the responses.
An additional problem interfering with establishing cause and effect relates to quantifying exposures in field and laboratory studies. To establish dose-response relationships and to compare the potencies of different effluents, effluent exposures typically have been quantified using percentage effluent. This measure of exposure alone does not take into account several factors that would influence the toxicological potency of an effluent. The types of wood species used, process type, water use, and efficiency of biotreatment all influence the potency of a given effluent, but these are not consistently reported. Bulk measures, such as BOD , total suspended solids , conductivity , polyphenolics , and measures of specific ions (Ca+ and K+) , frequently are used as supporting information to provide an estimate of effluent concentrations. Commonly measured effluent extractives, such as resin acids , plant sterols , and chlorophenolics , also have been used. Several studies do not report any measures or estimates of effluent concentrations [6,94,95]. These inconsistent measures of effluent quality throughout the literature have contributed to difficulties when interpreting bioassay responses between studies and needs to be addressed in future work aimed at finding solutions.
Toxicity source evaluation
To circumvent problems related to working with extremely complex final effluent matrices, several studies have adopted an approach that investigates individual in-plant process wastes as potential sources of reproductive effects. This TSE approach does not consider chemical modifications that can occur during biotreatment (see Role of effluent biotreatment) or from metabolic alterations, but it offers the advantage of eliminating wastes not involved in the effects as well as simplifying effluent matrices for potential TIE experiments. In one such study, goldfish were exposed to 13 wastewater streams from a bleached sulfite and a bleached kraft mill, and both circulating sex steroids and in vitro gonadal steroid production were measured . Except for the final biotreated BSME tested at 100%, none of the other wastewater streams caused statistically significant reductions. Similarly, only 100% effluent from the secondary clarifier of the biotreatment system elicited steroid responses with the wastewaters in the bleached kraft mill.
Using mummichog, a series of experiments have linked depressed steroids to condensates generated from weak black liquor evaporation during chemical recovery as well as from bleachery effluents [30,31,96,97]. This work has continued on-site at another Canadian kraft mill but with a different fish species, because the mill in question discharged into a freshwater recipient [68,69]. Bioassays of 21-d duration were conducted with fathead minnows exposed to 1 and 100% secondary-treated effluent and four process streams. Exposure to the final effluent caused ovipositor development and induction of VTG in males, whereas females developed male secondary sex characteristics. Two process streams were the dominant contributors to the effects measured in the final effluent. Exposure to the acid stream (24% of the total final effluent volume) produced an increase in egg production in females and in VTG induction in males. The alkaline stream (67% of the final effluent flow) decreased egg production and caused VTG induction and ovipositor development in males. Condensates at this mill also contained chemicals associated with the steroid depressions in mummichog . Additional studies are required to further delineate these sources.
Toxicity identification evaluation
In addition to TSE studies, attempts have been made to identify directly the substances responsible for reproductive effects. The chief method employed is bioassay-directed fractionation. One of the critical aspects in this approach is the choice of endpoint used to drive the chemical manipulations. The development of mechanistic-based bioassays with different pathways of fish reproductive perturbations and shorter turnaround times has enabled their use as investigative tools. Most of the work conducted to date has employed in vitro binding assays to characterize ligands for sex steroid receptors, chemicals associated with in vivo sex steroid depressions, and a combination of in vivo and in vitro techniques to evaluate androgenic and estrogenic substances.
A collection of international studies since 2001 has pursued the identities of compounds causing masculinization effects. Using in vitro tests, several attempts have been made to determine the cause of effluent-induced masculinization in Florida. Fenholloway River water receiving BKME was found to contain androgen-receptor agonists that were characterized as nonpolar organic material [6,99]. In one case, further fractionation and TIE work indicated androstenedione as the causative agent . An earlier study in which solid-phase extraction was followed by high-pressure liquid chromatographic fractionation, however, identified a fraction without androstenedione that caused androgenic activity . Further studies have highlighted the possibility that phytosterols can be converted to progesterone, and then to androstenedione and androstadienedione, by microbial populations in the water and sediment downstream from mills . Recent work from New Zealand, however, has shown that relatively high waterborne concentrations (10-100 μg/L) of androstenedione and androstadienedione are required to elicit masculinization in mosquito fish . This is consistent with other evidence from New Zealand that sterols present in effluents, as well as theirb oxidation products resulting from chlorine dioxide bleaching, are not involved in the masculinization of mosquito fish and did not show activity in rainbow trout brain androgen receptor– binding assays. Multitiered bioassay–directed fractionation experiments of primary and biotreated effluents from a chlorine-free Swedish kraft mill directed by androgen receptors isolated from Atlantic croaker (Micropogonias undulates) ovaries have shown that unidentified compounds possessing diterpenoid skeletons are present in multiple fractions exhibiting binding affinities and that a receptor-mediated pathway is the primary route by which masculinization effects seem to be occurring [8,84].
Using controlled fish exposures, a series of laboratory and field experiments have determined that Canadian BKME and BSME contain multiple ligands for fish sex steroid receptors and the aryl hydrocarbon receptor [37–39], demonstrating bio-accumulation of compounds capable of affecting steroid signaling after brief waterborne exposures. At the Canadian bleached kraft mill where condensates have been identified as a primary source of chemicals affecting steroid levels in mummichog, investigations also have shown that substances reducing testosterone are readily bioavailable, are polar, and can cause an effect within days . In-depth chemical characterization (solid-phase extraction, reverse-phase high-pressure liquid chromatography, and gas chromatography–mass spectrometry) of condensates have identified various lignin degradation products and terpenoids to be associated with the effects and to be present in condensates at another bleached kraft mill with documented effects on wild fish . The relevance of these findings to the broader industry (kraft and nonkraft processes as well as biotreated effluents) is the focus of current efforts.
To summarize, bioactive substances have been characterized as being polar as well as nonpolar, being water soluble as well as associated with solids, having the potential to act as ligands for a variety of receptors (including estrogen and androgen receptors), and mainly being nonbioaccumulative, requiring a sustained exposure to cause the effects. These findings further highlight the complexity of the presence of multiple bioactive substances functioning by multiple mechanisms. Questions remain about the possibility of common causative agents, particularly those related to wood furnish, that are independent of mill process type.
MILL PROCESS/OPERATING CONDITIONS AND EFFECTS ON FISH REPRODUCTION
Understanding how manufacturing processes, mill furnish, effluent biotreatment, and actual mill operating conditions (e.g., spill control) can influence effluent quality are important pieces of information for understanding the sources of the reproductive effects and developing corrective strategies. Throughout the course of studies regarding effluent effects on fish reproduction, the industry has undergone substantial changes in operating conditions. These include changes in bleaching practices to eliminate the formation of 2,3,7,8-tetrachlorodibenzodioxin, the increasing use of TMP and bleached chemithermomechanical pulping processes, and the installation and/or upgrading of secondary-treatment facilities. This provides a good opportunity to track the influence of these changes on effluent quality, but it is difficult to form generalizations about the mills of today based on studies conducted 10 years ago. Today, for example, virtually all mills have effluent biotreatment for meeting regulatory BOD and acute lethal toxicity limits. In the 1980s and early 1990s, however, only approximately 60% of the mills in Canada had effluent biotreatment . Nevertheless, reviewing the literature on how mill processes/operating conditions may affect fish reproduction could provide clues as to the origin of the fish responses.
Manufacturing processes and effluent quality
Mills produce pulp mainly by chemical (e.g., kraft and sulfite) or mechanical (e.g., TMP) means; multiprocess mills employ both processes. Because the effluent components from mills using different manufacturing processes can vary substantially , questions exist regarding whether manufacturing processes themselves influence the ability of an effluent to affect fish reproduction. Swedish field studies during the 1980s indicated that effects on fish in the vicinity of bleached kraft mills were greater than effects near unbleached kraft mills [103–106]. Coupled with the discovery around the same time of 2,3,7,8-tetrachlorodibenzodioxin in effluents discharged by mills employing chlorine for bleaching as well as a small number of studies at mill sites using other kinds of processes, this led to the speculation that effects on fish were caused exclusively by chlorine-containing compounds . As additional studies were conducted, however, a different picture began to emerge.
The initial studies at Jackfish Bay involved an examination of effects associated with BKME. To investigate the health of fish exposed to effluents from other types of mills, fish were examined in the recipients at seven Canadian mills in Ontario during the early 1990s [23,108]. Of the seven study sites, four received inputs from bleached kraft mills, one from a mainly unbleached kraft mill, one from a sulfite mill, and one from a multiprocess mill. At that time, only three of the effluents were biotreated. In general, gonad size and plasma hormone concentrations in effluent-exposed fish were decreased irrespective of mill process, suggesting a uniform response to differing production types.
Laboratory tests with fathead minnow have since confirmed that fish reproduction can be affected by effluents from mills using different manufacturing processes. Thus far, tests have been conducted in North America with bleached and unbleached kraft mill effluent [54,63,65,66,109–112], BSME [62,113,114], effluents from TMP mills [1,54,61], and effluents from multiprocess mills [1,54]. One series of experiments used the same protocol (adult fathead minnows and effluent exposure for 21 d) and allowed a direct comparison of 11 effluents [1,54]. Effluents from four TMP, four bleached kraft, and two multiprocess mills were tested at concentrations of 2 and 20% (v/v); one BKME also was tested at 40% (v/v). Except for the effluent from one multiprocess mill, the remaining 10 effluents affected at least one reproductive endpoint.
With the objective of comparing effects from bleached mills employing kraft and sulfite digestion, 21-d laboratory tests with rainbow trout showed that both BSME and BKME decreased pregnenolone concentrations in immature rainbow trout and increased VTG . Only the kraft effluent, however, caused a reduction in circulating testosterone concentrations. Additional comparisons included studies of oxidative stress indicators and circulating sex steroids . In this case, both effluents caused similar oxidative stress responses, but substantial differences were observed in endocrine-disrupting potential. The BSME reduced pregnenolone concentrations in females and caused VTG induction. The BKME caused increased circulating testosterone concentrations and no VTG induction, suggesting changes in effluent quality had occurred.
Perhaps the greatest numbers of comparisons between effluent quality and field-based effects can be found with the latest national assessment from the Canadian EEM program . Within that program, each mill conducts field surveys in three- to six-year cycles. Each cycle consists of an adult fish population component as well as a benthic invertebrate community component. By comparing the data from each mill using both tabular and meta-analysis formats, it was determined that the responses of both fish populations and benthic communities were quite similar between the second (1996-2000) and third (2000-2004) cycles of monitoring . As mentioned previously, the national average response pattern for fish in both cycles consisted of a decrease in gonad weight with increases in liver weight and body weight. These results have been interpreted as a form of nutrient enrichment overlaid with metabolic disruption; fish exposed to mill effluents have a disrupted allocation of resources to reproduction. The emergence of a consistent, national average response pattern (from >60 mills and >40 fish species/cycle) is strong evidence that indicates responses are not dependent on variables such as types of wood feedstock, dilution, production, and biotreatment. Sublethal toxicity testing (a component of the EEM program) offish (larval growth and survival), invertebrate (reproduction and survival), and algae (reproduction and growth inhibition) is conducted twice each year for each mill using freshwater and marine species. Analysis of the second cycle of monitoring data showed no meaningful relationship between sublethal toxicity and fish population or benthic invertebrate data , underscoring the need for tests that are predictive of field effects if progress toward identifying causes and solutions is to be realized (see Moving Forward).
Overall, data from individual studies at selected sites and the national monitoring program in Canada demonstrate that effluent-related effects on fish reproduction are not related to the type of manufacturing process (pulping and bleaching). This suggests the possibility of a common source of effects, such as the wood furnish or additives used by the mills. This also suggests that reproductive effects most likely are minimized by improving mill operations, particularly in terms of spill prevention and effluent treatment, either of specific waste streams or of combined mill effluent.
Process changes can provide a confounding factor when correlating effects with process type, but several studies have used process changes as an opportunity to conduct before-and-after studies to ascertain whether such changes may result in improved effluent quality. Field studies have been conducted to assess the consequences of upgrades in mill operating conditions along the coast of Sweden in the Baltic Sea  and in Canada at Jackfish Bay [3,5] and on the Moose River basin [5,44,116]. In addition, the masculinization effect on female mosquito fish that was first observed in Florida during the 1970s has been monitored continuously [94,117]. In all cases, a general improvement in the reproductive status of fish, although not to what can be considered full recovery, has been observed following changes in mill operating conditions. In some cases, reproductive effects have been reduced, and in others, substantially higher effluent concentrations were required to cause the effect. For example, after a Swedish bleached kraft mill at Norrsundet converted to 100% chlorine dioxide (i.e., ECF) bleaching, no effects on sexual maturation of perch were evident . Gonadosomatic indices (proportion of gonad weight to body weight) were smaller in both sexes before the mill's conversion and were lower only in females after the conversion. In addition, the effects on larval survival were lessened, but not completely eliminated. These improvements also were reflected in terms of greater diversity/abundance of fish near the effluent discharge . The improvements were credited to reduced chlorine-containing compounds, and the residual differences were attributed to non-chlorinated substances.
As noted previously, the mill at Jackfish Bay also converted to high chlorine dioxide substitution bleaching during the 1990s and installed a biotreatment system. With time, improvements in hormone levels and gonad sizes were observed. Studies from the late 1990s indicated that gonadal apoptosis (i.e., programmed cell death) may affect steroid biosynthetic capacity . Reductions in apoptosis corresponded to improvements in the gonadosomatic index of white sucker in the Moose River basin . The changes observed in wild fish corresponded to improvements seen with caged goldfish in terms of circulating sex hormones and in vitro gonadal steroid production . The improvements in turn appear to correspond to in-plant changes, particularly increased chlorine dioxide substitution bleaching, although a definitive linkage has not been established.
Similar improvements in steroid levels following mill process changes also were observed in fish collected from waters receiving effluent from the two mills in the Moose River basin. The improvements occurred in one case as the mill process was changed from a mechanical/sulfite process to TMP (possibly the only evidence that one process [in this case, TMP] can produce better-quality effluent than a mechanical/sulfite process) and in another case as a bleached kraft mill tightened liquor losses. Both mills also installed effluent biotreatment [4,5,44,116,120]. Therefore, it is difficult to attribute these changes to a specific change in operating conditions.
In Florida, the masculinization of wild female mosquito fish also was noted to be diminished following changes in mill operating conditions [94,117]. In a pre- and postconversion study, both the degree and the extent of masculinization were diminished after a suite of changes were made in mill operating conditions for compliance with the U.S. Cluster Rules . The mill in question operates both bleached and unbleached kraft lines. The changes included conversion to ECF bleaching, reduction of black liquor losses, condensate stripping, and improvement of effluent treatment. The benefit of these changes also was monitored in largemouth bass exposed to various effluent concentrations under controlled conditions. The changes in mill operating conditions did not eliminate effects on gonad size or sex steroid levels, but the concentrations required to cause these effects were increased from 20 to 40% (v/v) effluent.
The ability of effluents before and after modifications to affect fish reproduction in laboratory tests also has been determined, and the results largely mirror observations with wild fish. Life-cycle fathead minnow tests with BKME before upgrades had been shown to reduce egg production, increase the proportion of males, and delay sexual maturity at effluent concentrations of 2.5% or greater . After conversion to ECF bleaching, improved condensate handling and biotreatment (increased retention time in the aerated stabilization basin), installation of Lo-Solids™; (Andritz, Statteggerstrasse, Graz Austria) cooking, and reduction of black liquor losses, no reproductive effects were noted at 20% (v/v), the highest effluent concentration tested [109,122]. Other pre- and postconversion studies [64,110,111] also found improved reproductive performance by minnows (25% inhibitory concentration for egg production increased from ˜14 to ˜61%) after a mill converted from CEHDED (chlorine [C], alkaline extraction [E], hypochlorite [H], and chlorine dioxide [D]) bleaching to ECF bleaching and installed oxygen delignification. The mills in question made other changes in addition to bleaching, such as more efficient brownstock washing, and this was reflected by reduced components in the effluent from the pulping side of the operation [64,110,111]. Finally, a study of a BSME before and after changes were made showed that the final biotreated effluent caused steroid depression in goldfish before, but not after, the changes . In this example, mill changes included alteration in the impregnation cycles, SO2 charge, and temperature in the digester; increase in ClO2 substitution for bleaching from 60 to 65%; reduction in loss of solids from the bleachery; improved chip pile management; increased aeration in the aerated stabilization basin; increased ClO2 generator efficiency; and shutdown of the deinking plant.
Biological impacts in both mosquito fish and rainbow trout were observed to occur in relation to some known operational changes at the New Zealand multiprocess mill. The tested bleached kraft/TMP effluent demonstrated the ability to cause reduced gonad size associated with decreased plasma steroid hormones in female rainbow trout and masculinization in female mosquito fish [48,49,124] well after the installation of secondary treatment in 1972 and the conversion from elemental chlorine in 1998. Experiments conducted after the initial testing revealed disappearance of the previously observed mosquito fish and trout reproductive impacts [18,59,93]; however, a new pattern of response was observed in rainbow trout, wherein males rather than females showed reduced gonad size . Between the earlier and later studies, the major in-mill improvement was conversion of the kraft pulp mill screen room to closed cycle in 2000, with all filtrate sent to the recovery boiler. After this time, the ethoxyresorufin-O-deethylase induction previously observed in rainbow trout also disappeared, and differences were noted in effluent chemistry . In addition to the screen room change, a number of environmental improvements, such as improved maintenance and dredging in the treatment system, have been ongoing at this mill. These changes have led to continually improving effluent quality, particularly in terms of suspended solids and extractives .
It should be noted that the majority of the aforementioned examples for improved effluent quality are from situations in which mills made several simultaneous changes to their operating conditions. One of the most widely implemented process changes—namely, conversion from chlorine-based to chlorine dioxide bleaching—has received much attention because of the environmental benefit attributed to reduced loadings of organochlorine chemicals in final effluents. The environmental benefits of this are obvious. Nevertheless, it is not possible to attribute improved reproductive performance in fish to this modification—or to any other individual modification—because the mills in question also made additional changes in other aspects of their operations [18,64,110]. Aspects of another process change with intentional environmental benefits— namely, secondary effluent treatment—are discussed below (see Role of effluent biotreatment).
A series of studies that specifically assessed the role of a single process change involved condensate treatment by reverse osmosis (RO) at a Canadian bleached kraft mill [30,31]. At the outset of that work, the effluent at the mill was not biotreated. Before RO treatment of fifth-effect evaporator condensates, a 27-d exposure to combined mill effluent caused a reduction in plasma testosterone of both male and female mummichog, but no differences were observed in 11-ketotestosterone in males or in 17β-estradiol in females. In vitro gonad incubation tests also were conducted and reflected the trends in plasma testosterone and 17β-estradiol levels, although 11-ketotestosterone production was significantly increased. After the RO treatment was installed, final combined effluent did not cause a decrease in plasma testosterone after 30 d of exposure; rather, it caused an increase after 57 d of exposure. The situation was similar for 11-ketotestosterone in males. Some inconsistencies occurred in the in vitro production of steroids, but the RO treatment of condensates improved effluent quality at concentrations of relevance to the receiving environment (1%). Laboratory exposures of mummichog to the condensates (RO feed), the permeate produced by the RO system (reused in the mill and released in the final mill effluent), and the concentrate (removed from the effluent waste stream) have confirmed fifth-effect evaporator condensates to be the primary source of compounds causing hormone depressions in mummichog at this mill as well as the ability of the RO system to remove the active compounds .
Role of effluent biotreatment
The role of biotreatment on effluent quality in relation to reproductive effects in fish requires special consideration. In Canada, most mills that did not already have effluent biotreatment responded to revised effluent regulations for BOD and acute lethal toxicity in 1992  (http://www.dfo-mpo.gc.ca/oceans-habitat/habitat/policies-politique/act-acte_e.asp) by installing biotreatment. Effluent biotreatment typically reduces BOD by 90% or more, and it eliminates acute lethal toxicity to rainbow trout and Daphnia magna. Effluent biotreatment also produces a significant improvement in effluent quality, as determined by sublethal tests used for the Canadian regulatory EEM program [126,127]. These tests include a 7-d fish growth test (e.g., fathead minnow for mills discharging into fresh-waters), an invertebrate (e.g., Ceriodaphnia dubia) reproduction test, and an algal growth assay. It is important to note that effluent treatment was not designed to address effects on fish reproduction, but an important question following industry-wide installation was whether these benefits of biotreatment were reflected in improved fish reproductive performance.
Information required to answer this question can be obtained indirectly as well as directly, from studies designed specifically to evaluate the role of effluent biotreatment. The indirect evidence comes from field studies. Two field studies in North America demonstrated improvements at the fish community level after the installation of effluent biotreatment [92,128]. Munkittrick et al. [3,4,21] reviewed the data from North American and Swedish field studies conducted over a 10-year period. During that time, the mills that were studied made many changes in addition to the installation of biotreatment. After reviewing the evidence, those authors concluded that although improvements were noticeable, neither biotreatment nor the other modifications made by the mills completely eliminated reproductive responses. In the Swedish case, improvements in plasma steroids and gonad size were not seen before the installation of biotreatment in the fall of 1992 [129,130]. After the installation of biotreatment, the gonad size of perch was unaffected by effluent exposure until 1995 , but later studies indicated that perch from the effluent-exposed zones again had smaller gonadosomatic indices in 1997 and 1998 .
Direct evidence comes from laboratory/artificial stream studies that tested the quality of effluent from the same mill before and after biotreatment. Here, some examples of improvements could be related to biotreatment. In one study , brown trout were exposed to effluent from a TMP mill before and after biotreatment (anaerobic and aerobic). The plasma testosterone concentrations in males and females were decreased by both effluents (four-month exposure), although the magnitude of the reduction was smaller when the fish were exposed to the biotreated effluent. The levels of 17β-estradiol were unaffected in fish exposed to either effluent. The hatch-ability of brown trout eggs was improved when the adults were exposed to biotreated effluent from the TMP mill compared to untreated effluent . In a second study , female mosquito fish were exposed to an integrated bleached kraft/TMP mill effluent before and after biotreatment installation. Females were masculinized to a lesser extent following biotreatment, although some masculinization was still apparent. In a third study , biotreated effluents from both a bleached kraft mill and a TMP mill were found not to affect most reproductive endpoints measured in fathead minnows (egg production, fertilization, number of spawning events, percentage hatching, and plasma sex steroids), whereas effluent before treatment affected several of these parameters.
In other studies, no clear benefit was related to biotreatment because of cases where there were no difference in effects after biotreatment, cases in which some effects disappeared and others manifested, and finally, observations of more pronounced effects posttreatment. For example, in one experiment, goldfish were exposed to BSME and BKME . Before biotreatment (aerated stabilization basin), 10% BSME had no effect on production of testosterone and 11-ketotestosterone by male gonads (nonstimulated as well as stimulated by addition of human chorionic gonadotropin [hCG]) or on circulating levels of testosterone and 11-ketotestosterone . Higher concentrations caused fish mortality. In contrast, the biotreated effluent caused no mortality and was tested at 100% (v/v). A significant decrease in testosterone production by non-stimulated and hCG-stimulated testes and in circulating levels of steroids in blood was noted, although as described above, once the mill made a suite of changes (including biotreatment), testosterone production by hCG-stimulated testes was similar to that of controls. The latter finding suggests that at least after the modifications, biotreatment did not produce compounds with the potential to affect testosterone production by male gonads. The tests with BKME showed that before primary treatment, 40% concentration caused no reduction of 11-ketotestosterone production by the testes of goldfish, nor did 100% effluent exiting the aeration basin. Interestingly, 100% effluent after discharge from the secondary clarifier did reduce 11-ketotestosterone production by the gonads of male fish.
In another study , plasma testosterone levels in male and female mummichog were reduced by both 1% primary-treated and biotreated effluents from a bleached kraft mill, with the reduction being greater when the fish were exposed to the biotreated effluent. In this case, biotreatment caused an increase in the ability of effluent to depress fish steroids. Finally, when guppies (Poecilia reticulata) were exposed to effluent from a bleached kraft mill before and after biotreatment, evidence of masculinization (based on color) was reported irrespective of treatment . This did not occur in a concentration-dependent fashion, however. Effluent before biotreatment caused the effect at 5 and 25% concentrations, whereas after biotreatment, only fish exposed to 10% effluent were affected.
Based on the available information, the role of biotreatment on effluent quality in relation to potential effects on fish reproduction is inconclusive. The available data indicate that biotreatment may be beneficial in some cases but have little influence—or even worsen effects—in others. Relatively few studies, however, have made a direct comparison of effluent quality before and after biotreatment, and of those that have, several were confounded by incomparable test conditions. Thus, it is not known if the contradictions regarding biotreatment are a consequence of the different measures of reproductive performance that were studied (e.g., sex steroids and secondary sex characteristics), local conditions in field studies (e.g., fish receiving a higher effluent exposure, because biotreated effluent is nontoxic), or differences in biotreatment efficiency or other process changes made at the same time. Clearly, the role of biotreatment in the effects on fish reproduction needs to be studied further.
Based on a review of the literature, a substantial database regarding the effects of effluents from pulp/paper mills on fish reproduction is available. Clearly, a tremendous amount of research and monitoring have been conducted, and this has led to useful information and insights. Despite the enormous efforts and progress to date, no breakthroughs leading to solutions for these effects have occurred. Because of the complexities involved, a concerted effort by an experienced, multidisciplinary group ultimately will be necessary for solutions to be realized for the pulp and paper industry. After conducting the present review, we have identified the following priority research needs required to achieve this goal.
Tools for investigating cause and solutions
Causal investigations, including TIE and TSE work, require bioassay endpoints that can be done in a reasonable time frame. The most practical endpoints are those measured in laboratory tests or in shorter-term, on-site bioassays. Although numerous laboratory tools exist, there has been no consensus to date regarding which are most appropriate for use in TIE or TSE work, for tracking the outcome of technological changes, or for making a decision as to when an effluent has reached an environmentally acceptable quality. The regulatory EEM studies in Canada have identified metabolic disruption as a national response pattern that includes reduced gonad size. At present, a practical test for gonad size does not exist, and it is not known if other measures of fish reproduction (e.g., sex steroid levels and egg production) are related to gonad size or predictive of the reductions in gonad size observed in wild fish. An evaluation of available tests is necessary to link the responses across the different levels of biological organization involved before any meaningful TSE and TIE work can be conducted and treatment solutions properly evaluated. Such an evaluation also would serve as a means to determine the relevance of emerging genomics tools. This work should include appropriate measures of effluent quality so that the potencies of effluents can be normalized beyond comparing percentage effluent.
TSE and TIE investigations
The wide array of effects on fish reproduction demonstrates the complexity of the issue and raises questions regarding whether such changes are caused by a common agent or by different substances. Until this is clarified, it is not known if common solutions will be possible or if solutions will need to be found on a site-specific basis. A variety of effluent components originating from the wood furnish as well as their chemical products (through the manufacturing processes [e.g., kraft cooking]) and biodegradation products (as can occur during biotreatment) have been shown to affect fish reproduction. Nevertheless, their actual role in accounting for the effluent-related effects is unclear. To date, the role of the tree species used has been poorly examined in studies pertaining to reproductive effect. Both TSE and TIE investigations, using appropriate measures of fish reproduction, are necessary to clarify the actual sources of the effects of present-day effluents, the chemical compositions of which have changed since the 1990s. It is important to note that it may be possible to make progress toward treatment solutions without exact knowledge of the chemical structures of the causative agents.
Effluent biotreatment evaluation
Effluent biotreatment has many beneficial effects in terms of effluent quality, for which it was designed. Effluent biotreatment has been shown to be effective in reducing some effects on fish reproduction (e.g., minimizing masculinization), to exacerbate other effects (e.g., sex steroid levels), and to have no effect on others. Relatively few studies have made direct comparisons or adequate measures of effluent quality before and after biotreatment. A rigorous evaluation of the different types of biotreatment, using appropriate measures of fish reproduction, are necessary to clarify what role existing treatment systems are playing in the responses and if they could be modified or upgraded and, thereby, represent a possible solution.
Numerous international studies have demonstrated that pulp and paper mill effluents have the potential to affect fish reproduction. Some progress has been made in identifying the sources and causes of the reproductive effects in fish. These investigations have demonstrated complexities in effluent composition, fish reproduction, and the way in which effluents elicit effects. When examining whether a relationship between manufacturing parameters and fish reproduction exists, the evidence to date indicates that the effluent-related effects are not associated with a particular type of manufacturing process or biotreatment system but, rather, more likely dependent on how a mill operates in terms of spill control, optimized biotreatment, and handling of specific wastewater streams (e.g., condensates). This suggests a common origin of causative agents, such as those originating from wood or additives or their biodegradation products. Just as evidence exists that effects can be caused by effluents from mills using different processes, evidence also exists for the absence of effects in relation to effluents from other mills using the same processes. Developing a surrogate measure of effluent potency via a measure of effluent quality may go a long way toward determining why effluents from mills using the same processes have different potencies in terms of reproductive effects.
Although the rapid, mechanistically based bioassays employed in studies since the year 2000 have provided endpoints to direct TIE and TSE investigations, care must be taken when extrapolating the results of these experiments to effects observed in wild fish (e.g., depressions in gonad size). Linkages between in vitro and short term in vivo tests and effects observed in wild fish remain unclear and need to be established if the sources, identities, and ultimately, treatment solutions are to be found for the pulp and paper industry.