Aluminum chloride hydroxide sulfate

Long-term toxicity to fish

Link to relevant study record(s)

Reference

Reference 1

Endpoint:

fish early-life stage toxicity

Type of information:

migrated information: read-across based on grouping of substances (category approach)

Adequacy of study:

key study

Study period:

1989

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

other: Study is complete, method has undergone scientific peer review. raw data is presented. Only draw-back is no GLP, therefore a 2 rating

Qualifier:

according to guideline

Guideline:

other: Cleveland, L., E.E. Little, 8J. Hamilton, D.R. Buckler, and J.B. Hunn. "Interactive toxicity ofaluminum and increased acidity to early life stages ofbrook trout," Trans. Am. Fish. Soc. 115:610-620 (1986).

Qualifier:

according to guideline

Guideline:

other: Ingle, S.E., J.A. Keniston, and D.W. Schultz. "Aqueous chemical equilibrium computer program," Unites States Environmental Protection Agency. EPA-60013-80-049 (May 1980).

GLP compliance:

not specified

Analytical monitoring:

yes

Details on sampling:

Water samples were collected at the beginning of the exposures and weekly
thereafter to determine aluminum concentrations. Our analytical approach involved
separating the "fast reacting" or labile aluminum species (inorganic and organic) from
the "slow reacting" polymeric forms. A m.odification of the oxine/methyl isobutyl
ketone (MIBK) extraction procedure of Barnes16 was used to separate the two forms of
aluminum.

Vehicle:

no

Details on test solutions:

The following four steps were incorporated into the method: (1)
determination oftotal aluminum in acidified unfiltered samples (weekly determinations
for each treatment); (2) determination of total aluminum in filtered acidified samples
(two determinations per treatment); (3) determination of extr"dctable aluminum in
filtered samples (twodeterminations pertreatment); and (4) determination ofextractable
aluminum in unfiltered samples (3-4 determinations per treatment). The commercial
. in-line apparatus used to filter water samples contained 0.1 um mixed cellulose acet.ate
and nitrate filters. The filters were rinsed with 50 mL of 1% nitric acid and then with
50 mL of reverse osmosis (RO) water before use. Aluminum stock solutions were
prepared in RO water obtained from a commercial RO water purification system and
anionic, cationic, and mixed-bed exchange resins. In addition, the REDEQL-EPAK
chemical equilibrium model of Ingle et al., 1980 was used to estimate aluminum speciation
during the two exposures.
To determine filterability of aluminum at a nominal pH of 5.5, we used filters of
several types and pore sizes to filter 100-ml water samples that contained nominal
aluminum concentrations of 150 Jlg/L.
Atomic absorption standards (1000 ug/L: J.T. Baker Chemical Company,
Phillipsburg, NJ) were used for the preparation of standard solutions of aluminum. All
standards were preserved in 1% nitric acid solutions

Test organisms (species):

Salvelinus fontinalis

Details on test organisms:

The eyed eggs were obtained from Beity's Resort,
Valley, Washington. and held overnight in control water at pH 7.2 and 12°C before the study was initiated.

Test type:

semi-static

Water media type:

freshwater

Limit test:

no

Total exposure duration:

60 d

Nominal and measured concentrations:

nominal: 0, 38, 50, 75, 100, 150, 200, 300, 400 ug/L
measured: 5 (background), 4, 8, 29, 57, 68, 88, 142, 169, 292, 350 ug/L (total aluminium)

Details on test conditions:

We placed 50 eyed brook trout eggs into each of four 177-mL glass hatching containers, which were suspended in each
duplicate exposure tank. The tanks were covered to shield the eggs from light, and
aeration provided continuous circulation of the exposure water. Egg mortality, hatch,
and incomplete hatch were recorded dailyuntil hatching was complete. Dead eggs were
removed and discarded and newly hatched larvae were transferred to duplicate
containers within each exposure tank. Eggs yielding larvae that were unable to
complete the hatching process and had the chorion attached were considered to be
incompletely hatched. The median hatching date for each treatment was considered to
be day 0 of the larval exposure. On day 15, two groups of 20 larvae were randomly
selected from the hatching containers, photographed for length determination,14 and
placed in the smaller growth chambers within each duplicate tank. The remaining
larvae were placed in the larger chamber of each tank for behavioral, biochemical, and
aluminum bioconcentration measurements.
Mortality offish in the growth chambers was recorded daily and analyzed for days
15, 30,45 and 60 of the exposures. These fish were photographed again on day 30 for
length determinations. At day 30, each small chamber was thinned to 10 fish each to
avoid the effects of over-crowding on growth. Lengths and weights were detetmined
from direct measurements on days 45 and 60. Behavioral evaluation on 10 fish from each duplicate treatment included measurements of locomotory activity, feeding and
swimming capacity, and buoyancy of individual fish at 30 and 60 days of exposures A
and B. Frequency of movement was the number of times a fish changed position as it
began movementor changed directionduring a 2-min interval. Feedingwas the number
of strikes directed atprey-- Daphnia magna or Artemia salina-- during a 5-min period.
Swimming capacity was measured in a stamina tunnel as fish were subjected to
incremental increases in water velocity until fatigue occurred. The buoyancy of fish
was assessed by assigning a score of 3 to fish that remained on the aquaria bottoms, 2
to fish that swam in midwater, and 1 to fish that swam at the water surface,
Whole-body RNA and DNA were determined8 at 15 and 30 days of exposure Aand B.
Whole-body aluminum residues were determined on a wet-weight basis for 10 fish
per duplicate treatment on day 15 and 5 fish per duplicate on days 30, 45, and 60.15
Tissue residue data were statistically analyzed to derive surface response plots based on
whole-body aluminum residues, aluminum exposure concentrations. and days of
exposure.

Reference substance (positive control):

no

Duration:

60 d

Dose descriptor:

NOEC

Effect conc.:

14 µg/L

Nominal / measured:

meas. (geom. mean)

Conc. based on:

dissolved

Remarks:

Aluminium

Basis for effect:

other: swimming capacity

Remarks on result:

other: pH 5.6-5.7

Duration:

60 d

Dose descriptor:

NOEC

Effect conc.:

15 µg/L

Nominal / measured:

meas. (geom. mean)

Conc. based on:

dissolved

Remarks:

aluminium

Basis for effect:

weight

Remarks on result:

other: pH 5.6-5.7

Duration:

60 d

Dose descriptor:

NOEC

Effect conc.:

24 µg/L

Nominal / measured:

meas. (geom. mean)

Conc. based on:

dissolved

Remarks:

Aluminium

Basis for effect:

mortality

Remarks:

in Fry

Remarks on result:

other: pH 5.6-5.7

Duration:

60 d

Dose descriptor:

NOEC

Effect conc.:

26 µg/L

Nominal / measured:

meas. (geom. mean)

Conc. based on:

dissolved

Remarks:

Aluminium

Basis for effect:

mortality

Remarks:

in Fry

Remarks on result:

other: pH 6.5-6.6

Duration:

60 d

Dose descriptor:

NOEC

Effect conc.:

26 µg/L

Nominal / measured:

meas. (geom. mean)

Conc. based on:

dissolved

Remarks:

Aluminium

Basis for effect:

weight

Remarks on result:

other: pH 6.5-6.6

Duration:

60 d

Dose descriptor:

NOEC

Effect conc.:

22 µg/L

Nominal / measured:

meas. (geom. mean)

Conc. based on:

dissolved

Remarks:

Aluminium

Basis for effect:

other: Swimming capacity

Remarks on result:

other: pH 6.5-6.6

Duration:

60 d

Dose descriptor:

NOEC

Effect conc.:

13 µg/L

Nominal / measured:

meas. (geom. mean)

Conc. based on:

dissolved

Remarks:

Aluminium

Basis for effect:

number hatched

Remarks on result:

other: pH 6.5-6.6

Details on results:

Total dissolved aluminum measured in filtered acidified
samples ranged from 13 to 21 ug/L in exposure A and from 12 to 22 ug/L in exposure
B; total extractable aluminum in samples that were extracted immediately without
filtration ranged from 39 to 211 ug/L in exposure Aand from 27 to 194 ug/L in B; and
total dissolved aluminum in filtered extracted samples ranged from 14 to 27 ug/L in exposure A and from 13 to 26 µg/L in B. Results from the REDEQL-EPAK modeling suggest that a large portion of the total aluminum present in these studies was
in the solid, or undissolved, form. This was confirmed by the results obtained
from our filtration procedures which showed that concentrations ofdissolved aluminum
were consistently low. Filtration of exposure water with 3.0-um Teflon,
0.4 um polycarbonate, 0.2 um nylonand Fluoropore (Millipore Corporation, Bedford,
MA), and 0.1 um mixed cellulose-acetate and nitrate filters gave similar results. The
filtration results, coupled with the REDEQL-EPAK estimates, suggest the presence of
particulate or polymeric aluminum species larger than 3.0 um.

Reported statistics and error estimates:

One-way analysis of variance was used to compare treatment effects on survival,
growth, behavior, and nucleic acid content in both exposures. Percent data were arcsine
transformed, and square root transformations were made on strike frequency and water
column position data prior to statistical analyses. The Least Significant Difference
means comparison test was used to distinguish differences among treatment means (p
::;; 0.05). Swimming capacity data were also analyzed by simple regression to determine
the relation betweenswimming capacity and aluminum exposure concentration. Wholebody
residues ofaluminum were analyzed as a completely randomized design in which
the treatments were arranged factorially. The linear statistical model contained the
effects of aluminum exposure concentrations, days of exposure, and interactions of
exposure concentrations and days of exposure. Linear and quadratic polynomial
orthogonal contrasts were computed to produce surface response plots for whole-body
residues of aluminum. Statistical Analysis Systems20 programs were used to perform
aU analyses on the mainframe computer system of the University of Missouri,
Columbia, Missouri.

Variable	pH 5.6-5.7		pH 6.5-6.6
	30d	60d	30d	60d
Movement/2min	68(15)	142(24)	169(26)	>350 (22)
Strike frequentie	8(-)	142 (24)	4d(-)	>350 (22)
Buoyancy	68(15)	--	169(26)	--
Swimming capacity	>292 (27)	29 (14)	88(26)	>350 (26)
Incomplete hatch^a	8b(-)	--	57(13)	--
Fry mortality	142 (24)	142(24)	169 (26)	169 (26)
Length	142(24)	142 (24)	57(13)	88(26)
Weight	142c (24)	68 (15)	88c (26)	88(26)
RNA Content	142 (24)	--	>350 (22)	--
DNA Content	>292 (27)	--	>350 (22)	--

Note: Values represent total aluminumand(in parentheses) total dissolved monomeric aluminum

aEmbryo responses were determined after all embryos had hatched or died.

bBackground aluminum measured in pH 5.7 control treatment

cDetennined at 45 days of exposure.

Validity criteria fulfilled:

not applicable

Conclusions:

The lowest NOEC was determined to be 13 ug/L (measured dissolved Al) for both the endpoint " incomplete hachting". The test is valid and the method has gone trough peer review. Therefore the data can be used in a risk assessment as key data.

Executive summary:

Two partial life-cycle studies were conducted for 60 days in a modified flow-through proportional diluter.n In the first exposure (exposure A), eyed eggs of brook trout and the resultant larvae and juveniles were exposed to nominal aluminum concentrations of 38, 75, 150, and 300 ug/L at pH5.5; controls (no aluminum added) were held at pH 5.5 and 7.2. In the second exposure (exposure B), eyed eggs and the resultant larvae and juveniles were exposed to nominal aluminum concentrations of 50, 100,200, and 400 ug/L at pH 6.5; controls were held at pH 6.5 and 7.2. Because aluminum is less soluble under less acidic conditions and its toxicity to fish decreases, we used slightly higher aluminum exposure concentrations in exposure B than in A. The exposures were conducted in soft water containing about 3.0 mg/L calcium. Aluminum sulfate hexadecahydrate (98% pure: 8.56% allminum) was purchased from Fisher Scientific Company, Fairlawn, New Jersey. The eyed eggs were obtained from Beity's Resort, Valley, Washington. and held overnight in control water at pH 7.2 and 12°C before the study was initiated.

The lowest NOEC was determined for incomplete hachting to be 13 ug/L dissolved Al.

Reference 2

Endpoint:

fish short-term toxicity test on embryo and sac-fry stages

Type of information:

migrated information: read-across based on grouping of substances (category approach)

Adequacy of study:

key study

Study period:

23-11-1988 to 18-04-1989

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

other: see 'Remark'

Remarks:

Although no data according GLP, study is complete and was performed with natural water with natural Al concentrations. Data on disturbing factors is partial missing, like feeding data. however when looking at the difference between the 28 and 42 day endpoints this argument is considered as having no significant impact on the reliability. Data on the method for mortality score is not clear, however, this may be considered as a minor flaw. Endpoints are usable for risk assessments, eventhough this is no GLP study. Therefore it is rated as reliable with restrictions.

Qualifier:

according to guideline

Guideline:

other: Stoner, J.H, A S Gee and K.R. Wade, 1984. The effects of acldlficaUon on the ecology of streams in the Upper Tywl catchment m West Wales Environ Pollut Ser A 35, 125-157.

Qualifier:

according to guideline

Guideline:

other: Whitehead, P G, S Bird, M Hornung, J. Cosby, C. Neal and P. Paricos, 1988 Stream acidificaUon trends in the Welsh uplands - a modelhng study of the Llyn Brmnne catchments. J. Hydrology 101, 191-212.

Qualifier:

according to guideline

Guideline:

other: Dnscoll, C.T, J P. Baker, J.J. Blsogm and C L. Schofield, 1980. Effect of ahimimum speciation on fish m dilute acidified waters. Nature 284, 161-164.

Qualifier:

according to guideline

Guideline:

other: Goenaga, X and D.J.A Wflhams, 1988. Alumimum specmtion in surface waters from a Welsh upland area. Environ. Pollut. 52, 131-149

Qualifier:

according to guideline

Guideline:

other: Finney, D.J, 1971. Problt analysis. Cambridge University Press.

GLP compliance:

not specified

Analytical monitoring:

yes

Details on sampling:

monomeric aluminium (Alm) was measured in samples pressure filtered through a 0.015 um membrane
filter; organic monomeric aluminium (Alom) was the fraction of a filtered sample
which passed through a cation exchange column of Amberlite resin (preconditioned
for pH and conductivity); inorganic monomeric aluminium (Al1) was Alm-Alom and
adsorbed aluminium (Alad) was Altm-Alm.

Vehicle:

no

Details on test solutions:

Natural water according to Stoner et al., 1984 and Whitehead et al., 1988

Test organisms (species):

Salmo trutta

Details on test organisms:

To ensure comparability, all life stages were derived from the same strain of brown
trout, and reared in the hatchery (pH 6.2, 22 mg CaCO31-1) until required. Hatchery
temperatures were about 3°C warmer than those in the study streams and fish held
here developed faster. This led to some overlap in the exposure periods of different
life stages.
eggs:
On 23 November 1988, eggs and milt were stripped from the parent fish and transported
separately to the study site. In order to maximise fertilisation success in the
field, milt and eggs were mixed dry in a plastic tray and left for 2 min. The ova were
divided into six batches, each placed in 5 1 of water from one of the six study streams.
After leaving to harden for 1 h, 300 fertilized ova from each batch were divided equally
amongst six egg boxes.
alevins:
Alevin experiments started on the 25 January 1989, approximately 17 days after
hatching.

Test type:

other: natural streams

Water media type:

freshwater

Limit test:

no

Hardness:

< 20 mg CaCO3/L

pH:

4.7-6.9

Nominal and measured concentrations:

measured Al monomeric concentrations: 3, 5.2, 5.6, 12, 34, 56, 88, 377, 397 ug/L

Details on test conditions:

eggs:
The ova were
divided into six batches, each placed in 5 1 of water from one of the six study streams.
After leaving to harden for 1 h, 300 fertilized ova from each batch were divided equally
amongst six egg boxes. Egg boxes were cylindrical, 13 cm high by 7 cm diameter,
and constructed from 2 mm plastic mesh reinforced with 5 mm mesh. They were filled
with gravel 1-3 cm in diameter in order to reduce egg contact and siltation. Boxes
were buried 8-10 cm deep (to the box centre) in artificial redds located in gravels
typical of those used for natural spawning.
alevin:
Alevin cages were of similar construction to the egg boxes, measuring 20
cm high by 12 cm diameter. These were filled with approximately 16 cm depth of
gravel and placed in buckets of stream water. Eighteen alevins were added to each
cage and allowed several minutes to disperse into the gravel. The cages were then
buried in similar locations to the eggs, leaving the top 4 cm exposed. This allowed
for the emergence of the 'swim-up' stage. Six cages, three for both pre- and post-
'swim-up' alevins were used in each of nine streams
Parr:
Parr of age 80 days were held in 10 streams for a maximum exposure
period of 21 days In each stream, four cages (2 mm
mesh) each holding five fry were secured to the bed by steel stakes. Survival was recorded
every 3-4 days. The limitation in exposure time was imposed to minimise any
confounding stress through reduced ability to feed.

Reference substance (positive control):

no

Duration:

28 d

Dose descriptor:

LC50

Effect conc.:

19 µg/L

Nominal / measured:

meas. (geom. mean)

Conc. based on:

dissolved

Remarks:

aluminium

Basis for effect:

mortality

Remarks:

pH 5.8-5.9 effects on Alevins

Duration:

42 d

Dose descriptor:

LC50

Effect conc.:

15 µg/L

Nominal / measured:

meas. (geom. mean)

Conc. based on:

dissolved

Remarks:

Aluminium

Basis for effect:

mortality

Remarks:

pH 5.8-5.9 effects on alevins

Details on results:

Apart from LI6, survival of eggs to hatching ranged from 71.1% to 84.0% (means
back-transformed from arcsine values), and was apparently independent of variations
in aluminium concentration between streams.
Mean percentage alevin mortality increased with aluminium concentration and exposure
time.

Exploratory regression of mortality probits (Finney, 1971) on mean concentrations of each aluminium species (log transformed), indicated that mortality in both 28- and 42-day exposures was most strongly related to mean total monomeric aluminium (Altm). For alevins exposed for 28 days the following regression explained 81.3 % of the variance in mortality, mortality probit = 1.24 + 2.93 x log[Altm], giving an LC50 of 19 ug/L Altm. For 42 day exposure alevins ('swim-up'), 94.2% of the variance in mortality was explained by the regression equation, mortality probit = 1.35 + 3.12 x log[Altm], with an LC50 of 15 ug/L Altm. LC50 values for Al-filterable were 79 ug/L for 28 -day alevins and 72 ug/L for 42 -day alevins, although this fraction accounted respectively for only 54.8 % and 37.7 % of the variance in mortality. Alevin mortality could also be related to mean pH for both exposure periods. For younger alevins the regression was, mortality probit = 21.5 - 2.85 x pH. (r^2 = 97.4 %) For 'swim-up' alevins, mortality probit = 20.3-2.61 x pH. (r^2= 85.2%) Approximate LC50 values were pH 5.8 and pH 5.9, respectively. Survival of alevins in the limed streams was high, as expected from the low aluminium concentrations and high pH at these sites.

Validity criteria fulfilled:

not applicable

Conclusions:

The long-term LC50s were determined to be 19 and 15 ug/L for 28 and 42 days of exposure in natural waters, respectively. Values are measured values of total monomeric Al. Study is valid and scientific sound. although essential data are missing, these data should not have a impact on the actual endpoint and are therefore considered not significant enough to have an impact on the reliablility. The endpionts can therefore be used in a risk assessment especially when lokking at acidic waters.

Executive summary:

The survival of artificially implanted eggs, alevlns and parr of brown trout Salmo trutta (L.) was assessed m streams of different acidity. Chemical analysis included detailed alumlnium speciatmn of surface and interstitial water samples, taken over the duration of lntragravel hfe stages. Egg survival, from two minutes after fertilization to hatctung, was usually above 71%, and was independent of the mean concentration of total monomeric aluminium (Altm) over the range 3 -397 ug/L The survival of alevins exposed for 28 days (before 'swim-up') or 42 days ('swim-up') was most strongly related to mean Aim concentration and to pH For 28- and 42-day exposures, LC50 values for Altm were approximately 19 and 15 ug/L, respectively, or 79 and 72 ug/L for 0.45 gm filterable alumlnium (Al-filterable) The 21-day LC50 of parr ca 3 mth old, was between 84 and 105 ug/L mean Al-filerable concentration. During a simulated acid episode of mean pH 4 8 and 880 ug Al/L, parr showed 100 % mortality after only 18 h. We discuss the problems of quantifying toxic responses in soft-water streams. As exemplified in this study, chemical variables show wide and rapid fluctuations, and are often co-correlated These features confuse interpretation of field experiments or predlctmn from laboratory data. The reviewer considers the most applicable endpoint to be the LC50 regarding total dissolved monomeric Al to alevin of 19 and 15 ug/L.

Description of key information

EPA-60013 -80 -049, key study, validity 2 (Cleveland, 1989):

60 days-NOEC (Salvelinus fontinalis) = 13 µg/L (based on geometric mean of measured concentrations)

Key value for chemical safety assessment

Fresh water fish

Fresh water fish

Effect concentration:

13 µg/L

Additional information

Two read-across studies performed on aluminium sulphate are available to determine the long-term toxicity to fish of the registered substance :

1) Two partial life-cycle studies were conducted for 60 days in a modified flow-through proportional diluter. In the first exposure (exposure A), eyed eggs of brook trout and the resultant larvae and juveniles were exposed to nominal aluminum concentrations of 38, 75, 150, and 300 ug/L at pH5.5; controls (no aluminum added) were held at pH 5.5 and 7.2. In the second exposure (exposure B), eyed eggs and the resultant larvae and juveniles were exposed to nominal aluminum concentrations of 50, 100,200, and 400 ug/L at pH 6.5; controls were held at pH 6.5 and 7.2. Because aluminum is less soluble under less acidic conditions and its toxicity to fish decreases, we used slightly higher aluminum exposure concentrations in exposure B than in A. The exposures were conducted in soft water containing about 3.0 mg/L calcium. Aluminum sulfate hexadecahydrate (98% pure: 8.56% allminum) was purchased from Fisher Scientific Company, Fairlawn, New Jersey. The eyed eggs were obtained from Beity's Resort, Valley, Washington. and held overnight in control water at pH 7.2 and 12°C before the study was initiated.

The lowest NOEC was determined for incomplete hachting to be 13 ug/L dissolved Al.

2)The survival of artificially implanted eggs, alevlns and parr of brown troutSalmo trutta(L.) was assessed m streams of different acidity. Chemical analysis included detailed alumlnium speciatmn of surface and interstitial water samples, taken over the duration of lntragravel hfe stages. Egg survival, from two minutes after fertilization to hatctung, was usually above 71%, and was independent of the mean concentration of total monomeric aluminium (Altm) over the range 3 -397 ug/L The survival of alevins exposed for 28 days (before 'swim-up') or 42 days ('swim-up') was most strongly related to mean Aim concentration and to pH For 28- and 42-day exposures, LC50 values for Altm were approximately 19 and 15 ug/L, respectively, or 79 and 72 ug/L for 0.45 gm filterable alumlnium (Al-filterable) The 21-day LC50 of parr ca 3 mth old, was between 84 and 105 ug/L mean Al-filerable concentration. During a simulated acid episode of mean pH 4 8 and 880 ug Al/L, parr showed 100 % mortality after only 18 h. We discuss the problems of quantifying toxic responses in soft-water streams. As exemplified in this study, chemical variables show wide and rapid fluctuations, and are often co-correlated These features confuse interpretation of field experiments or predlctmn from laboratory data. The reviewer considers the most applicable endpoint to be the LC50 regarding total dissolved monomeric Al to alevin of 19 and 15 ug/L