QUANTITATIVE STRUCTURE-ACTIVITY RELATIONSHIPS FOR CHRONIC TOXlClTY OF PHENOL, P-CHLOROPHENOL, 2,4-DICHLOROPHENOL, PENTACHLOROPHENOL,

QUANTITATIVE STRUCTURE-ACTIVITY RELATIONSHIPS FOR CHRONIC TOXlClTY  OF PHENOL, P-CHLOROPHENOL, 2,4-DICHLOROPHENOL, PENTACHLOROPHENOL,
Scientific Excellence Resource Protection & Conservation Benefits for Canadians
Excellence scientifique Protection et conservation des ressources Bénéfices aux Canadiens
4
QUANTITATIVE STRUCTURE-ACTIVITY
RELATIONSHIPS FOR CHRONIC TOXlClTY OF
PHENOL, P-CHLOROPHENOL,
2,4-DICHLOROPHENOL, PENTACHLOROPHENOL,
P-NITROPHENOL AND 1,2,4=TRICHLOROBENZENE
TO EARLY LlFE STAGES OF RAINBOW TROUT
(Oncorhynchus mykiss)
Hodson, P.V., R. Parisella, B. Blunt, B. Gray, and K.L.E. Kaiser
Physical and Chemical Sciences Branch
Department of Fisheries and Oceans
Maurice Lamontagne Institute
Box 1000, Mont-Joli (Québec) G5H 324
Canadian Technical Report
of Fisheries and Aquatic Sciences 1784
1+1
Fisheries
and Oceans
Pêches
et Océans
=
I
I
Canadian Technical Report of
Fisheries and Aquatic Sciences
Technical reports contain scientific and technical information that contributes to
existing knowledge but which is not normally appropriate for primary literature. Technical
reports are directed primarily toward a worldwide audience and have an international
distribution. No restriction is placed on subject matter and the series reflects the broad
interests and policies of the Department of Fisheries and Oceans, namely, fisheries and
aquatic sciences.
Technical reports may be cited as full publications. The correct citation appears
above the abstract of each report. Each report is abstracted in Aquatic Sciences and
Fisheries Abstracts and indexed in the Department's annual index to scientific and
technical publications.
Number 1-456 in this series were issued as Technical Reports of the Fisheries
Research Board of Canada. Numbers 457-714 were issued as Department of the
Environment, Fisheries and Marine Service, Research and Development Directorate
Technical Reports. Numbers 715-924 were issued as Department of Fisheries and the
Environment, Fisheries and Marine Service Technical Reports. The current series name
was changed with report number 925.
Technical reports are produced regionally but are numbered nationally. Requests
for individual reports will be filled by the issuing establishment listed on the front cover
and title page. Out-of-stock reports will be supplied for a fee by commercial agents.
Rapport technique canadien des
sciences halieutiques et aquatiques
Les rapports techniques contiennent des renseignements scientifiques et techniques
qui constituent une contribution aux connaissances actuelles, mais qui ne sont pas
normalement appropriés pour la publication dans un journal scientifique. Les rapports
techniques sont destinés essentiellement à un public international et ils sont distribués à
cet échelon. Il n'y a aucune restriction quant au sujet; de fait, la série reflète la vaste
gamme des intérêts et des politiques du ministère des Pêches et des Océans, c'est-à-dire
les sciences halieutiques et aquatiques.
Les rapports techniques peuvent être cités comme des publications complètes. Le
titre exact paraît au-dessus du résumé de chaque rapport. Les rapports techniques sont
résumés dans la revue Résumés des sciences aquatiques et halieutiques, et ils sont classés dans l'index annuel des publications scientifiques et techniques du Ministère.
Les numéros 1 à 456 de cette série ont été publiés à titre de rapports techniques de
l'Office des recherches sur les pêcheries du Canada. Les numéros 457 à 7 14 sont parus à
titre de rapports techniques de la Direction générale de la recherche et du développement, Service des pêches et de la mer, ministère de l'Environnement. Les numéros 715 à
924 ont été publiés à titre de rapports techniques du Service des pêches et de la mer,
ministère des Pêches et de l'Environnement. Le nom actuel de la série a été établi lors de
la parution du numéro 925.
Les rapports techniques sont produits à l'échelon régional mais numérotés à l'échelon national. Les demandes de rapports seront satisfaites par l'établissement auteur dont
le nom figure sur la couverture et la page du titre. Les rapports épuisés seront fournis
contre rétribution par des agents commerciaux.
Canadian Technical Report
of Fisheries and Aquatic Sciences 1784
Quantitative Structure-Activity Relationships for
chronic toxicity of phenol, p-chlorophenol,
2,4iiichlorophenol, pentachlorophenol,
p-nitrophenol, and 1,2,4-trichlorobenzene
to earPy life stages of rainbow trout
( Oncorhynchus mykiss )
P.V. Hodsona, R. Parisellat, B. Blmt3, B. Gray3 and K.L.E. Kaiser4
Physical and Chemical Sciences
Department of Fisheries and ûceans
Maurice Lamontagne 1nstitute
C.P. 1000
Mont-Joli (Québec)
G5H 324
Maurice Lamontagne Institute, 850,
Mont-Joli, Québec, G5H 324.
route de
la mer, C.P.
1000,
Pulp and Paper Research Institute of Canada, 570 St. John's Boulevard,
Pointe-Claire, Québec, H9R 3J9.
Great Lakes Laboratory of Fisheries and Aquatic Sciences, Canada
Centre for Inland Waters, Box 5050, Burlington, Ontario, L7R 4A6.
National Water Research Institute, Canada Centre for Inland Waters,
Box 5050, Burlington, Ontario, L7R 4A6.
Minister of Supply and Services Canada 1991
Cat. No. Fs 97-6/1784E
ISSN 0706-6457
Correct citation for this publication:
Hodson, P.V., R. Parisella, B.R. Blunt, B. Gray and K.L.E. Kaiser. 1990.
Quantitative Structure-Activity Relationships for chronic toxicity of
phenol,
p-chlorophenol ,
2,4-dichlorophenol,
pentachlorophenol,
p-nitrophenol and 1,2,4-trichlorobenzene to early life stages of rainbow
Can. Tech. Rept. Fish. Aquat. Çci. 1784:
trout (Oncorhynchus mykiss).
55 p .
TABLE OF CONTENTS (CONTsD)
.....................
...................
......................
2. 4.Dichlorophenol
1.2. 4.Trichlorobenzene
Pentachlorophenol
.
.
.
........
EFF'ECT THRESHOLD ESTIMATES
.
.
.
.
Reqreçsion estimation
NOEC-LOEC geometric mean
Bootstrap estimation
Summary
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...........
...........
...........
...........
...........
LIST O F TABLES
Pase
Table
....
5
.......................
7
1
P r o p e r t i e s and a c u t e t o x i c i t y of t h e test chemicals.
2
Range of c o n d i t i o n s i n exposure tanks throughout t h e
toxicitytests
.
12/13
...
16/17
3
Mean c o n t r o l responses of rainbow t r o u t t o t e s t compounds.
4
Mean responses of rainbow t r o u t t o phenol exposure and
r e s u l t s of a n a l y s i s of variance and Dunnettfs t e s t . .
5
6
7
8
Mean responses of rainbow t r o u t t o p-nitrophenol exposure
and r e s u l t s of analysis of variance and Dunnett's t e s t
...
Mean responses of rainbow t r o u t t o p-chlorophenol exposure
and r e s u l t s of analysis of variance and Dunnett's t e s t . .
.
24/25
Mean responses of rainbow t r o u t t o 2,4-dichlorophenol
exposure and r e s u l t s of a n a l y s i s of variance and
Dunnett'stest...........
............
26/27
Mean responses of rainbow t r o u t t o 1,2,4-trichlorobenzene
expssure and r e s u l t s of a n a l y s i s of variance and
meétsa%est.
32/33
..a...................
9
10
11
20,21
Mean responses of rainbow t r o u t t o pentachlorophenol
exposwe and r e s u l t s of a n a l y s i s of variance . . . .
....
35
Linear r e g r e s s i o n s r e l a t i n g measurements t o loqarithm of
concentration ( y = a t b log X). Thresholds a r e i n t e r p o l a t e d
values a s s o c i a t e d with a 25% change r e l a t i v e t o c o n t r o l
measurements . . . . . . . . . . . . . . . . . . . . . . . . .
38/39
Threshold concentrations associated with a 25% change
r e l a t i v e t o c o n t r o l values . . . . . . . . . . . . . .
40,41
...
LIST OF TABLES (CONT'D)
Paqe
Table
12
13
14
Comparison of threshold estimates based on larval or
fry weight reduction and mortality increase.
44
Comparison of acute and chronic toxicity of test compounds
torainbowtrout....
47
Relationship between log P and the logarithms of acute and
chnonic toxicity estimates (in uM) of the test chernicals
48
........
..................
..
LIST OF FIC;URES
Pase
Fiqure
1
2
The temperature regime and duration of each life stage durinq
rainbow trout embryo-larval-fry tests.
...........
The effect of exposure to phenol on the wet weiqht of
rainbow trout larvae at hatch. Threshold effect
concentrations are s h o w for each statistical method of
estimation
.........................
3
The effect of exposuse to p-nitrophenol on the wet weight
of rainbow tnout fry after four weeks of feeding. Threshold
effect concentrations are shown for each statistical method
of estimation.
.......................
4
The effect of exposure to p-chlorophenol on the wet weiqht
of "larvae onlylf, 10 days after hatch. Threshold effect
concentrations are shown for each statistical method of
estimation
.........................
5
The effect of exposure to 2,4-dichlorophenol on the total
mortality of rainbow trout four weeks after swim-up.
Threshold effect concentrations are s h o w for each
statistical method of estimation
..............
6
The effect of exposure to 2,4-dichlorophenol on the wet
weight of fry four weeks after swim-up. Threshold effect
concentrations are ahown for each statistical method of
estimation except the geometric mean
............
7
8
The effect of exposure to 1,2,4-trichlorobenzene on wet
weight of rainbow trout four weeks after swim-up.
Threshold effect concentrations are shown for each method
of estimation. . . . . . . . . . . . . . . . . . . . . . .
.
The effect of exposure to pentachlorophenol on the wet
weight of rainbow trout at four weeks after swim-up. The
threshold effect concentration was estimated £rom the
regression . . . . . . . . . . . . . . . . . . . . . . . .
.
3
LIST OF FIGURES ( O N T ' D )
9
The relationship between octanol-water partition coefficient
and measures of acute and chronic toxicity for phenol,
p-nitrophenol, p-chlorophenol, 2,4-dichlorophenol,
1,2,4-trichlorobenzene and pentachlorophenol
........
46
Rainbow trout were expssed to vaterborne phenol, p-chlorophenel,
2,4-dichlorophenolt p-nitrophenol or 1,2,4-trichlorobenzene for 85 days. This
priod included full egg development from the day of fertilization, plus
htching, yolk resorption and four veeks of feeding as freely-swimming fry.
The primary effects of exposure were to reduce grovth rate and to increase
mortality rate. Grovth inhibition vas the most sensitive response since it
oecurred at exposure levels equal to or Power than those that inereased
iarsrtality rates. Changes in development rate, grovth efficiency and percent
msàsture were also observed after exposure to some chemicals, Changes within
experbments due to treatment effects were generally larqer than observed
variations of control responses compred amonq expriments.
Threshold
exposure concentrations for the chronic toxicity of each chemical were
ealculated £rom reqressions of the responses against the logarithm of exposure
concentrations. The threshold was that concentration predicted to change a
response by 25% relative to the control responses. The order of toxicity
bsed on these thresholds vas phenol < p-nitrophenol < p-chlorophenol
< 2,4-dichlorophenol < 1,2,4-trichloro-benzene.
Thresholds for pentachlorophenol were also calculated from a published study of chronic toxicity to give
a larger data base.
These data defined a Quantitative Structure-Activity
Relationship (QSAR) èetween the loqarithm of threshold effect concentrations
and the loqarithm of the octanol-water partition coefficient. This QSAR was
parallel to a QSAR for the acute lethality of these same chemicals. Since the
ratio of the slopes of the two QSARs waa 0.10, the data sugqests that there is
a constant ratio of about 0.10 between chronic and acute toxicity.
Des truites arcs-en-ciel ont été exposées pendant 85 jours au phénol, pchlorophénol, 2,4-dichlorophénol, p-nitrophénol ou 1,2,4-trichlorobenzène.
Pendant cette période, il y avait succesivement le développement des oeufs,
l'éclosion, la résorption du vitellin, et l'alimentation des jeunes pendant
quatre semaines avec une nourriture artificielle. Les effets principaux
étaient une rMuction du taux de croissance et une augmentation du taux de
mortalité.
L'inhibition de croissance était la réponse la plus sensible,
parce qu'elle arrivait h des niveaux d'exposition équivalents ou inférieure h
ceux qui changeaient les taux de mortalité. Quelquefois, il y'avait d'autres
réponses comme un changement de la vitesse de développement, de l'efficacité
de croissance, et du pourcentage d'humidité d a m les tissus. En genéral, les
changements dus aux prduits chimiques étaient plus grands que les variations
normales mesurées entre les expériences. Les seuils d'exposition associés
avec la toxicité chronique ont été calculés par des régressions entre les
réponses et les logarithmes des concentrations de chaque composé chimique
testé. Les seuils étaient les concentrations estimées pouvant induire un
changement de 25 % relativement h la réponse du témoin. L'ordre de toxicité
indiqué pas les seuils était phénol < p-nitrophénol < p-chlorophénol < 2 , 4 dichlorophénol < 1,2,4-trichlorobenz&nee
k plus, pour avoir une base de
données plus large, les seuils pour les réponses au pentachlorophénol ont ét&
calculés A partir d'une étude déjh publiée. Ces données ont defini une
Relation Quantitative entre la Structure et la Toxicité (RQST), c 'est-hdire, entre le logarithme des seuils de concentration et le logarithme du
coefficient de distribution des composées chimiques entre le n-octanol et
l'eau. Ce RQST était paralléle A une RQST calculée pour la toxicité létale
aique des mêmes prcduits chimiques. Les deux RQSTs suqgérent qu'il y a un
rapport constant de 0.10 entre la toxicité chronique et la toxicité aigue,
parce que le rapport entre les pentes était de 0.10.
An important part of water
quality management is the protection of aquatic species £rom toxic
chemicals, particularly £rom the
insidious effects of chronic exposures. Water quality criteria have
been established to define thoae
levels of chemical exposure that
are "aafeV, based on literature
reviews of controlled laboratory
experiments. However, for most of
the estimated 70,000 chemicals in
cornmon use, aquatic toxicity data
are limited to acute toxicity tests
(van Leuwen et a l . 19901. Data
£rom full life-cycle tests with
long-lived species are even more
scarce due to the high cost of
extended tests. Since many organic
chemicals
already
contaminate
aquatic food chains (Hesselberg
and Seelye 19821, extensive impacts
may occur before environmental
limits can be established.
life-cycle tests (Macek and Sleight
1977; McKim 1977), although their
accuracy is quite variable (Suter
1990).
Chronic toxicity and criteria
to limit chemical levels in water
have often been estimated by
ffApplicationFactors".
These are
fixed ratios between chronic and
acute tsxicity (usually 0.1 or
0.01), based on the assumption that
the ratio is invariate. While this
approach was acceptable in the past
when data were limited (e.g. NAS/
NAE 1972), modern practice demands
a higher level of proof. Furthermore there is no guarantee of a
fixed relationship between acute
and chronic toxicity because:
(a)
chronic toxicity may arise £rom
different mechanisms of toxicity;
(b) metabolic detoxification can be
induced during chronic exposure;
and (c) acute toxicity is often
measured near the limits of
solubility.
The need to rapidly identify
potentially hazardous chemicals and
to protect sensitive species has
engendered a variety of strategies,
including
truncated
chronic
toxicity tests (Macek and Sleight
1977; McKim 1977) and predictive
models of toxicity. Models may be
based
on ratios of acute and
chronic toxicity or on Quantitative
Structure-Activity
Relationships
(QSARs, Veith et al. 1983).
QSARs
compare
measured
toxicity to structural or physicochemical
properties
(molecular
descriptors) of the test chemicals.
QSARs based on tests of related
chemicals
with
a systematic
variation in molecular descriptors
can predict the toxicities of other
related but untested chemicals,
thereby limiting testing effort.
Truncated chronic studies test
cmbryogenesis or larval development, often the most sensitive
stage of an organismfs life-cycle.
Hence early life stage studies may
predict
toxicity during full
QSARs have been successfully
developed for the acute toxicity of
chemicals to yeast (chloroanilines,
Kwasniewska and Kaiser 19831, algae
(azaarenes, aromatic amines and
nitroaromatics, Schultz and Moulton
19841,
Daphnia
(hydrocarbons,
chlorinated hydrocarbons, oils,
Bobra et al. 1984) and various
species of fish (phenols, chlorophenols,
aliphatics, benzenes,
1981; Ve ith
narcotics, Konemnn
et al. 1983).
The most uçeful
molecular descriptor
was the
octanol-water partition coefficient
(P) which describes the relative
distribution of chernicals between
water and n-octanol. The slopes of
QSARa are very similar among
studies, indicating
that the
characteristics most relevant to
toxicity are chernical solubility
and ability to penetrate lipid
membranes
Measures of log P
effectively mimic the partitioning
of chemicals £rom water into lipids
of aquatic biota.
.
Although embryo/larval studies
are less expensive than full
life-cycle
tests, they still
require long testing times; there
are few QSARs based on truncated
tests
(anilines, chlorobenzenes,
van Leuwen et al. 19901. If the
potential of QSARs
is to be
realized and successfully applied
to chemical control, we must test
the assumption that QSARs based on
acute toxicity psedict chronic
toxicity.
We measured the toxicity to
embryos, larvae and fry of rainbow
trout (Oncorhynchus mykiss) of a
group of arorratic compounds chosen
to give a wide range of water
solubilities, partition coefficients, and acute toxicities.
Est imated
threshold
effect
concentrations were used to develop
a QSAR with log P. We tested the
hypothesis of a fixed relationship
between acute (96 hour LC50s) and
chronic toxicity by comparing the.
slopes of their respective QSARs.
Finally, since a standard test
protocol for rainbow trout was not
available,
we
examined
the
variability of control responses
among toxicity tests.
MATERIALS AND METHODS
DES1GN
Rainbow
trout eggs were
exposed to six concentrations
(including control) of each test
ehemisal: phenol (PH); p-nitropheno1 (NP); p-chlorophenol (Cl?); 2,4
dichlorophenol (DCP) ; and 1,2,4,
trichlorobenzene
(TCB).
The
concentration range spanned an
order of magnitude £rom 0.05 to 0.5
Increments of
of the 96-hr LC50.
concentration vere equal on a
logarithm scale, on the assumption
that each successive increment of
response requires a doubling of the
exposure intensity, according to
Beerfs law. The duration of each
test was £rom the day of fertilization, through hatch and yolk sac
resorption, to
four weeks of
feeding, a total of 85 days (Figure
1).
The life stages exposed were
(egqs) , larvae ( newlyembryos
hatched fish with obvious yollc
sacs - also known as alevins or
sac-fry), and fry (freely-swimminq
juveniles that feed exogenously).
Larvae that have absorbed their
28 Days
Larvae
( Alevin )
( Sac Fry 9
-
35 Days
HATCH
DAYS
Figure 1. The temperature regime and duration of
rainbow trout embryo-larval-fry tests.
each life stage during
yslk sacs and t b t leave the bottom
to begin feeding are called swim-up
fry. Temperatures were maintained
a% lOoC for egg developïnent, 120C
for yolk resorption and 150C for
fry growth (Figure 1)
.
Each experiment was replicated
three times for a total of 18
tanks. Each replicate within any
expriment received the fertilized
eggs £rom a different female, and
the order of chemical concentrations was randomized within each 6
tank replicate
Measurements of
toxicity included hatching success,
mortality, rate of development,
prevalence of de£ormities and
growth.
.
The test
chemicals were
either purchased as highly purified
standards sr repurified according
to the methmis listed (Table 11,
The test chemicals were added
to water directly (TCB)
or as
acpeous solutions (PH, NP, Cg, DB)
by a microlitre syringe p m p
linked to a diluter (Mount and
Brungs 1967). An electronic timing
device controlled both the pump and
a solenoid valve supplying water to
the diluter.
The dilutions were set for O
(control), 10, 18, 32, 56 and 100
percent of the maximum concentration. The flow ranged £rom 300 to
600 rnL per
minute
at each
concentration, and the output of
each ce11 was split three wayç
before distribution to replicate
exposure tanks. The exposure tanks
were glaas aquaria measuring about
40 cm long by 20 cm wide by 20 cm
high for a total volume of about
16 L, maintained at 14 L by a
standpipe.
Flow rate per tank
ranged from 125 to 225 mL per
minute, giving a 95% molecular
replacement time of 3 to 5.5 hours
(Sprague 1973).
At a desired
loading rate of 0.5 g of fish per
Pitre of water per day, the f low
limited total biorrrass per tank to
about 90 to 162 g.
The
concentrations
of
eikemlcals at each expoçure lewl
were rmieasured daily by UV absorpspectrophotometry
using
tion
standards prepared daily in control
aquarium water. Levels of TCB were
gas-liquid
determined
chromatography
after
solvent
extraction.
Al1 experimnts used water
from Lake Ontario, delivered to the
lab as a dechlori~tedmunicipal
supply.
Hardness was 135 mg/L as
CaC03 and alkalinity 80 mq/L as
CaCOs .
Flow rates, dissolved
oxygen,
pH
and conductivity
measured during the experiments are
summarized in Table 2. Detailed
analyses of other characteristics
are reported in Hodson et al.
(1980).
Prüperties aiid acute toxicity of the test iherriicals.
WATER
LOG PARTITION
COEFF ICIENTa
LC50 FOR
RAINBOM TROUTb
LOT
(ieN)
!LOG Pl
(UN)
SOURCEc
NUNBER
...............................................................................................................................................
CHEHICAL
REFINING
HOLECULAR
WEIGHT
SOlUBILITY
94.1
'il00Od
1.49
103
BDH
07541
p-nitrophenol
159.1
36"
1.91
57
A
HE5518HE
Froi etherhexane
p-ihlorùphenül
128.6
210'
2.42
14.8
BDH
2450950
From hexane
16.0
E
A4A
Fros hexane
A
CE5814
phenol
8.3"
TECHNIQUE
---
Redistillation
Li1
I
pentachlarophenol
266.4
0,036d
5.12
0.6
F
31145-382
...............................................................................................................................................
a From Hansih and Leu, 1979.
b FYOIDHodson et al. 1984.
c A, Aldrich Che~nicalCoapany; BDH, BDH Cheaicals; C, Caiedon laboratorie~;E, Eastman Kodak; FI Fluka Chenical Corporation.
d Fior Jone;, 1981.
e Versihueren, 1983.
f From Suntio et al. 1388.
g From Yalkowski et al. 1379.
h
Frùw
US EPA, 1'380,
Froi hexanetoluene
FISH
Rainbow t r o u t eggs £rom t h r e e
females were purchased f o r each
test £rom Goosen's Rainbow Ranch,
ûtterville, Ontario.
After f e r t i l i z a t i o n , t h e y were taken t o t h e
lab
in
thermos
jugs and t h e
chemical exposure s t a r t e d . About
200-300
eggs
were a l l o c a t e d
i m p a r t i a l l y t o each exposure tank
f o r a t o t a l of 3500-4500 eggs per
expriment.
Each of t h e s i x tanks
w i t h i n a r e p l i c a t e received eggs
£rom t h e same a d u l é female.
Within each tank, eggs were
held i n k i t c h e n s i e v e s with nylon
s c r e e n bottoms and s o l i d s i d e s .
These
containers
encircled the
standpipe d r a i n
so
t h a t water
l e a v i n g t h e tank f lowed up through
t h e eggs b e f o r e s p i l l i n g over t h e
A screen
t o p of t h e s t a n d p i p e .
over t h e s t a n d p i p e prevented t h e
l o s s of l a r v a e a f t e r hatch. Daily
records were kept of dead eggs,
s s r t e d aecordinq t o whether eye
pigments were v i s i b l e (eyed e g g s ) ,
and
whether
mortality
occured
d u r i n g t h e process of hatching.
Temperature was recorded d a i l y i n
each tank
and cumulative t o t a l
t e m p e r a t m e s were used t o estimate
degree-days t o h a t c h ( t h e sum of
d a i l y temperatures).
The number of l a r v a e hatched
each day were counted and t r a n s f e r red £rom t h e egg baskets t o t h e
main tank t o keep a record of d a i l y
and cumulative hatch.
The number
of degree-days t o 50% hatch and
£rom hatch t o swim-up was estimated
by p r o b i t a n a l y ç i s t o g i v e a n index
of t h e r a t e of development.
The
number of deformed l a r v a e was a l s o
counted
and
expressed
as
a
percentaqe of
t h e t o t a l number
hatched
.
Yolk sac conversion e f f i c i e n c y
was measured on a day when more
t h a n 20 l a r v a e hatched.
Ten were
s e g r e g a t e d and
ten
were immed i a t e l y d i s s e c t e d t o measure t h e
s e p a r a t e wet and d r y weights of t h e
pooled l a r v a e and yolk sacs. Dry
weights were measured by d r y i n q t o
a c o n s t a n t weight a t 60 C i n a
convection oven. The remaining t e n
f i s h developed f o r a f u r t h e r t e n
days i n a s e p a r a t e egg i n c u i a t i o n
b a s k e t before k i n g d i s s e c t e d a s
imïicated.
Yolk sac conversion
e f f i c i e n c y was c a l c u l a t e d as t h e
r a t i o of weiqht g a i n of t h e l a r v a e
t o weight l o s s of t h e yolk d u r i n g
t h e 10-day period (Hodson and Blunt
1986).
Percent moisture of t h e
d i s s e c t e d l a r v a e was a l s o c a l c u l a t e d as a n index of osmoregulation
and c a t a b o l i s m .
When t h e yolk of each f i s h w a s
almsst completely used and t h e f i s h
began t o s w i m towards t h e s u r f a c e ,
was
initiated
with a
feeding
salmonid "starter" d i e t ( M a r t i n ' s
Feed
Mills,
Elmira,
Ontario).
Feeding was
done g r a d u a l l y and
carefully
to
minimize wastage;
feces
and
uneaten
food
were
removed d a i l y .
When feeding w a s
s u c c e s s f u l , t h e f i s h were counted
and weighed l i v e i n water. This
e s t a b l i s h e d t h e i n i t i a l weight and
biomass of feeding f r y . Weighing
w a s repeated a t weekly i n t e r v a l s .
When
total
biomass of a tank
approached t h e recommended l o a d i n g
r a t e s , 50% of t h e f i s h were removed
and d i s c a r d e d . Four weeks a f t e r
TABLE 2
Ranue of conditions in emosure tanks throushout the toxicitv tests.
Chemical
Flow
(ml/min 1
OWgen
( mg/L
PH
Conductivity
( Fimhos/m2
phenol
196-206
11.0-11.5
7.97-8.02
244-245
p-nitrophenol
160-196
9.4-10.9
7.83-7.96
244-245
p-chlorophenol
195-205
11.6-11.8
8.08-8.10
242-243
225
8.6-11.0
7.87-8.08
267-283
pentachlorophenol
"
-.,..."
"
-.--.-....----
feeding began, ten fish £rom each
tank were removed to measure pooled
wet and dry weights and percent
moisture.
Mortality of fish (both larvae
and fry combined) was expressed as
a percent of the total in each tank
at hatch. To avoid bias caused by
removing fish, the total number of
fish-days (cumulative number of
fish in any tank on each day) was
also calculated and mortality
expressed as the number per 1000
fish-days
.
As eggs within each replicate
were derived
£rom a different
femle, the responses of fish to
chemical exposure were tested by a
randomized complete block design
for ANOVA (Steel and Torrie 1960).
A two-tailed Dunnett's test (Steel
and Torrie 1960) determined which
treatment means were different £rom
control. For unequal sample sizes,
we used a Dunnettfs test with the
Kromer modification (Day and Quinn
1989). The minimum difference £rom
control that could be declared
significant by Dunnett's test is a
funetion of the measured variance
and was calculated
for each
parameter.
STATISTICÇ
Changes in each measurement
were assessed by analyses of
variance
(ANOVA) and post-hoc
comparisons of treatment means with
controls. Threshold concentrations
for chemical effects were estimated
f rom
exposure-response
curves
defined by
linear regression
analysis
QSARs for chronic
toxicity
were
developed
by
comparing threshold exposure levels
to
octanol-water
partition
coefficients by linear regression.
Al1 tests were at the 0.05
probability level.
.
Prior to ANOVA, Shapiro-Wilk's
normality test, Bartlett's homogeneity of variance test, plots of
residuals, and plots of means vs
variances were used to identify
serious violations of assumptions
about normality and homogeneity of
variance (Sokal and Rohlf 1981).
Statistically
significant
responses may occur coincidentally
at least once in every 20 tests at
the 0.05 probability level. Since
30 tests were carried out per
experiment there was a high risk of
'Ifalse positives".
To ,minimize
this risk, we did not consider a
response significant unless it met
1) a
the following criteria:
significant "F" statistic for
treatments £rom the ANOVA; 2) a
monotonic
exposure-response
relationship; isolated or irregular
sesponses in the middle of an
exposure range were not accepted;
and 3) a change £rom the control
response by 25% or more.
For responses meeting these
criteria, the threshold effect
concentration was defined as that
concentration associated with a 25%
change £rom control values, as
suggested by the US EPA (19891.
The threshold concentration was
estimted by three methods. The
first calculated the geometric mean
of the LOEC, defined as the Lowest
Observed
Effect
Concentration
causing a change £rom the control
response of 25% or more, and the
NOM=, the next lowest or No
Observed Effect Concentration. The
NOM: and LOM: de£ine a range
equivalent to the MATC (Maximum
Acceptable Toxicant Concentration,
McKim 1985). The threshold is the
geometric mean os the mid-point of
this range on a log scale.
The second method calculated
the
simple linear regression
relating responses to the logarithm
of exposure concentrations. The
threshold was interpolated as that
concentration associated with a 25%
change in response £rom control
values.
Parameters were tested
only if they met the following
criteria: 1) çignificant treatment
effects demonstrated by the ANOVA;
2) largest response at least 25%
greater than control; 3 ) and.
responses at a minimum of three
exposure levels.
The third technique was linear
interpolation by the "Bootstrap
Methai" (Norbert-King 1988; US EPA
1989 1 .
Al1 ca1cuPation.s were
performed by the ffBOOTSTRP I Q V
software package
(developed by
Battelle Columbus, in cooperation
witk the US EPA 1989). This method
assumes
that
responses
are
monotonically decreasing (the mean
response for each higher concentration is less than or equal to the
mean response for the previous
concentration) and that responses
follow a piece-wise linear response
function (a linear response £rom
one concentration to the next).
If
means
do
not decrease
rnonotonically, the responses are
wsmoothedt' by averaging adjacent
means.
QSAR ANALYSIS
A chronic QSAR was developed
by simple linear regression of log
threshold concentrations determined
by the regression method versus
log P. An acute lethality QSAR was
similarly developed
£rom the
published LC5Os for these chemicals
for rainbow trout (Hodson et al.
1984).
The homogeneity of QSAR
slopes were compared by a t-test
(Sokal and Rohlf 1981).
The pentachlorophenol ( P B )
data in this report were derived
£rom a previous study (Hodson and
Blunt 1981).
The methodology of
the P B expriment was the prototype for the current work, but the
temperature regime (120, 150 and
200 for egg development, yolk sac
resorption and fry growth respectively), number of treatments (4) and
number of replicates (2) differ
£rom the current protocol. The PCP
data were the response means and
summary statistics £rom a factorial
ANOVA that tested concentration,
temperature
and concentrationtemperature interaction effects
(P 5 0.01).
No post-hoc cornparisons of
treatment means with controls were
possible .
However LOEC values
could be estimated for responses
which
exhibited
significant
concentration effects, and vhich
changed only at the highest
exposure level.
The 25 percent
effect threshold concentrations for
the P B data were interpolated £rom
linear
regressions
relating
response to the logarithm of
exposure concentration.
WATER CONDITIONS
The routine water analpes
demonstrated that water quality in
remained constant
eaeh
test
throughout the tests.
There were
no differences among treatment and
control conditions, other than in
the
level of added chemicals
(Table 2).
The levels of added
chemicals fluctuated slightly and
separate average concentrations are
reported for each phase (eggs,
larvae, fry) of each test to ensure
accurate estimation of thresholds
(Tables 4-8 ) . Spectrophotometric
determination of 8 and CU? proved
difficult due to interferences.
Gas liquid chromatography, after
solvent extraction and derivitization, indicated that measured
concentrations
were close to
nominal.
However, frequent
determinations proved too time
consuminq and costly. Therefore,
reported concentrations for 8 and
ï?CP (Tables 6 and 7) are nominal
values.
Daily records indicated that
water temperatures throughout al1
tests remaineà within loC of target
values.
Mean
control
responses
(Table 3) indicated a considerable
variation among experiments in the
percent mortality of al1 eggs,
which ranged £rom 2.5 to 24.6%.
This range appears to reflect
variability in egg fertilization
among experiments, as m n y eqqs
were unfertilized (no obvious ce11
division as s h o w by microscopic
examination, lack of cloudiness,
occurrence of large yellow oil
droplets, no movement), and should
not be considered true mortalities.
Eyeà eggs, those that developed
sufficiently
to
be
easily
recognized as living, showed a much
smaller range of control mortality,
£rom O to 10.4%.
One control
replicate in the CP expriment
exhibited a 60% mortality for eyed
eggs, compared 'to less than 3% for
the remaining two replicates. The
same replicate had a 75% mortality
for fry.
Tank contamination was
suspected
and
this
control
replicate was removed £rom the
analyses.
Percent
mortality
during
hatch was also much more consistent
(0-7.3%). The inverse of percent
mortality, percent hatch (i.e. the
percentage of eggs surviving to
hatch), showed the same ranges of
variability on a total and eyed
egg basis as percent mortality of
eggs.
Average development rates
for egqs varied from 340 to 3 8 1
degree-days,
corresponding
to
average temperatures during egq
development of 9.8
to 9.60C;
development rate increased ( fewer
degree-days)
with
average
incubation temperature.
At hatch, the weights of
whole fish varied £rom 67.6 to
81.4 mg among experiments (Table
3),
with a standard deviation
equivalent to 7.8% of the overall
control mean. When considered by
components, the larvae alone also
had a relative standard deviation
of 7.8%, with the yolks accountinq
for the missinq weight and havinq a
standard
deviation
of
8.7%.
Percent moisture was not as
variable (SD = 1.4%).
Percent mortality of sac fry
and development rate were quite
consistent (Table 3) but data for
NP and DB were missing. Percent
deformities (bug eye, bEue sac,
corkscrew) were somewhat more
variable, with a standard deviation
of 5.9%.
At 10 days post-hatch, the
average weight of whole fiah varied
£rom 83.2 to 107.2 mg amonq expriments (Table 3); the standard
deviation was 10.3%, compared to
7.8%
at hatch.
However, the
weiqhts of larvae alone at 10 days
post-hatch and the weights of yolk
had standard deviations of 16.7%
and 33.0% respectively, compared to
their hatch weight counterparts of
7.8% and 8.7%. Increasing variability indicates possible differences
rates and growth
in growth
efficiencies among experiments.
The conversion efficiency of yolk
exhibited a standard deviation of
20.2%Percent moisture varied
Pittle except for DB, which had a
mean value of 75%, about 12% lower
than that of
larvae in other
experiments
Within-group variations for the wet and dry weiqhts
indicated that dry weiqht was
overestimated
in one replicate,
which biased
the mean percent
moisture.
.
During
feedinq,
percent
msrtality of fish ranged £rom 6.6%
to 20.0% or 1.97 to 5.00 mortalities per 1000 fish-days (Table 3).
These values represent both larval
and fry mortality combined. Total
nwiber of fish-days varied considerably (range:
4375 to 9653),
demonstratinq a standard deviation
26.4% of the overall control mean.
Fry weiqht at 4 weeks post swim-up
also differed considerably amorig
exper iments
wi th
a standard
deviation of
27.7%.
We ight
differences for fry did not follow
the same relative order as larvae
due to differences in qrowth rates
among experiments.
Overall,
variability
in
measured parameters increased with
the stage of development as
different rates of development were
express&.
RESULTS OF ANALYSIS OF VARIANCE AND
DUNNETT'S TEST
Since each expriment included
m n y measurements, only those that
showed changes are discussed in the
text, but al1 results are reported
in the tables. A change is defined
Mean control respun~esuf rainbov trout t o t e s t coipounds.
...................................................
3
PH
tiortality üf al1 eggs
Hortality of eyed eggs
Hatch of eyed eggs
Z Died while hatching
1 Hatch of al1 eggs
Degree days tlj hatih
Wet weight of "larvae ünlyhaf hatch
Dry weight of 'lauvie onlye a t hatch
Wet weight of yolk sac a t hatch Img)
Dry weight of yolk sac a t hatch h g )
#et weight of larvae and yolk sac a t
Dry weight
larvae and yolk sac a t
Z fiüisture at hatch
lüf
11.6
Isgb
(mg)
hatch (mg)
hatch (mg)
LARVAE
1 tiortality of larvae
Z Deformities
Degree-days from hatch tu suii-up
#et weight o f ullarvae o n l y 5 a t 10 days (mg)
Dry weight of "larvae onlyu aat 10 days (mg)
Wet weight of yolk s a ï a t 10 days (mg)
Dry weiqht of yolk sac a t 10 days (mg)
Wet weight of larvae and yolk sac a t 10 days (mg)
Dry weight of larvae and yolk sac a t 10 days (mg)
Yülk sac iùnversiün efficiency (wet)
Yolk sac conversion efficiency iday)
Z Hüistuïe a t 10 days
4.6
95.4
3.5
88.4
340.0
27.1
3.8
41.8
19.0
69.0
22.9
85.8
3
NP
24. O
3.0
97.0
2. O
76.0
381.0
24.6
2.8
43. O
17.7
67.6
22.5
88.0
2
CP
3
DCP
3
TCB
tiean
Standard
Deviation
Range
TABLE 3 iiüntinued!
Nean c ü n t r ü l responses of rainbou t r u u t t o test cümpü~nds.
3
PH
I M ü r t a l i t y üf fish
M o r t a l i t y per 1000 fish-days
Total number of fish-days
Met weight a t week 4 post swia-up (mg)
Dry weight a t week 4 post suim-up (mg)
I Hüisture a t week 4 post swim-up
7.9,
1.97
7090
320
61.7
80.8
3
NP
20.0
5.00
7281
452
87.4
81.0
.
2
CP
3
DCP
15.2
3.40
9653
191
95.4
6.6
2.00
6958
701
80.6
-----
3
TCB
Mean
13.1
3.71
4375
539
110
73.5
12.6
3.22
7071
50 1
88.6
80.5
PH, phenol; NP, p-nitrophenül; CP, p-chlorophenol; DCP, 2,4-diihll:lrophen1jl; TCB, 1,2,1-trich1orl:lbenzene.
U-88
no d a t a .
Standard
Deviation
5.5
1.3
1870
140
20.2
0.7
Range
6.6 1.97 4375 320 61.7 79.5 -
20.0
5.00
9653
704
110
81.0
as a statistically significant
difference among the treatment
means.
Phenol
No effects of PH on eqgs were
observed during development but one
test replicate, for the 30 uM level
of PH, had a mortality rate of 97%
for eggs due to a sudden rise in
tank water temperature. The number
of replicates at this exposure
level was reduced to two (Table 4 ) .
At
hatch,
signif icant
decreases were observed in the mean
weight of larvae-onlyftrelative to
Decreases followed an
control.
exposure-response trend with a 39%
reduction in weiqht at the highest
exposure concentration (Figure 2).
Although the weights of whole fish
( larva
plus
yolk )
decreased
relative to
control (a 148
reduction at the highest test
concentration), none were statistically significant. At the highest
test
concentration,
percent
moisture at hatch
increased a
maximum of 2% relative to the
Larvae also
control mean.
demonstrated a significant increase
in percent mortality (Table 4 ) , by
about 17% relative to control at
the highest test concentration.
ff
Weiqhts of PH-exposed larvae
decreased significantly at 10 days
post-hatch. Decreases followed an
exposure-response
relationship
terminating in a 44% reduction in
weiqht relative
to controls.
Althouqh yolk sacs at 10 days
post-hatch were about 26% larger
than controls at the highest test
concentration, the change was not
siqnificant due to large withingroup variations.
Siqnificant
reductions in the weights of whole
fish (larva plus yolk) were
observed, with the mean weiqht at
the highest test concentration 25%
lower than control.
A 24% reduction in yolk sac
conversion efficiency was observed
at the hiqhest test concentration,
but responses were erratic: two
responses at mid-range concentrations increased relative to
control and changes were not
1n
statistically significant.
contrast, percent moisture of
10 days
post-hatch
larvae at
increased by only 1% relative to
control and
the change was
statistically significant.
Fry percent mortality followed
an
obvious
exposure-response
relationship (Table 4 ) terminating
in a 65% increase in mortality
compared to control. The corresponding increase in mortality per
1000 fish-days was £rom 2.0 to
25.4.
The mean weiqht of fry
decreased by about 27% at the
second hiqhest PH concentration,
but was only 15% lower than control
at the hiqhest concentration.
These changes were not statistically siqnif icant due to large
but
within-group
variations,
percent
moisture
iricreased
siqnificantly by 3% at the highest
test concentration.
Y = -5.79 LOG X + 27.17
R
B
G
C
O. 4
0.8
=
=
=
=
(
n = 5 , r2= 0 . 7 9 )
Log regresrlon threshold estimate
Log bootatrap threrhold estlmade
Log geometrlc mean threshold estlmate
Control response
1.2
1.6
LOG PHENOL CONCENTRATION ( p M )
Figure 2.
The effect of exposure to phenol on the wet weight of rainbow trout
larvae at hatch. Threshold effect concentrations are shown for
each statistical method of estimation.
Mean responses o f rainbow trout to phenol exposure and results of analysis o f variance and Dunnett's test.
3
3
3
3
2
3
S
s,
F
D
X
dif f .
phenï~l concentratiun ( y H j
1 Hvrtality of eyed eggs
Z Hatch of eyed eggs
t Died while hatihing
Degree days to hatch
Wet weight ut "larvai inly' a t hatch
Dry weight oi "larvae only" a t hatch
Wet weight o f yolk s a ï a t hatch (mg)
Dry weight of yolk sac a t hatch (ag)
Met veight of larvae and yolk sac a t
Dry weight of larvae and yolk sac a t
Z Moisture et hatih
(mg)
(mg)
hatch (ag)
hatch (mg!
0.0
4.6
95.4
3.5
340.2
27.1
3.8
41.8
19.0
69.0
22.9
85.8
LARVAE
phenol concentration (pli)
0.0
Z Murtality ü f larvae
3.0
1 Defurnied
5.1
Degree days fram hatct~ to swia-up
101.2
#et weight of "larvae only" a t 10 days i i g )
67.7
Dry weight of "lafvae u n 1 y " a t 10 days (mg)
9.5
Wet weight of yolk sac a t 10 days (mg)
24.2
Dry weight of yolk sac a t 10 days (mg)
12.1
Wet weight of larvae and yolk sac a t 10 days (mg) 91.9
Dry Wight of larvae and yolk sac a t 10 days (ag) 21.7
Yülk sac conversion efficiency (wet)
2.31
Yolk sac conversion effiiiency (dry)
0.83
Z l o i s t u r e a t 10 days
85.9
4.84
3.1
96.4
1.4
347.0
22.5
2.9t
38.5
17.4
60.8
20.8
87.4
8.46
2.7
97.3
0.5
341.2
22.9
2.9~
40.2
18.6
63.1
21.5
87.2
16.00
3,é
96.4
0.6
348.6
19.5~
2.6,
39.6
18.4
59.1
21.0
86. 6
30.20
2.8
97.2
0.3
352.4
20.0*
2.3
40.0
18. 6
60.0
20.9
88.5*
TABLE 4 iiüntini~ed!
Hean respünses üf rainbüw trout to phenol exposure and results of analysis of variance and Dunnett's
3
3
3
3
2
test.
3
S
!id
F
D
...............................................................................................................................................
Z
diff.
FRY
phenül concentration )!yi
X lortality of fish
Hortality per 10OO fish-days
Tutal nuiber o f fish-days
Net weight at week 4 po5t swim-up (mg)
Dry weiyht at week 1 püst swii-up (mg)
Z Hüisture at week 4 püst swis-up
0.0
2.93
5.21
7.9
9.2
13.7
1.97
2.17
3.60
7090
10179
8072
319.7
311.5
302.8
61.7
58.7
55.7
80.8
81.1
81.6
13.60
15.1
4.00
8083
311.3
57.9
81.4
29.40
42.30
37.M
73.01
8.32
6.79
10.35W
25.401 1.61
1.31
7353
1624 1131
1070
234.0
271.1
93.90 76.72
41.5
45.7
18.60 15.19
83.4,
0.55
0.45
82.2
27.48 21.59
95.80 4.16
5.67 3403
0.29
0.47
8.15
1.43
22
48
1
standard deviatiür~ calculated from poüled error variance; s,, standard difference between means used to calculate D value for Dunnettls test;
FI F value frur ANOVA; D, ninimum significant difference from control in original units as calculated by Dunnett's test; X d i f f . , minimum
significant percent difference from control.
Ç,
r Treatwent neans significantly different fros control by Dunnettls test.
For
most
responses, NP
'exposure had no effects on eggs
(Table 5), but differences were
observed for developmental rate.
Two
concentrations
increased
degree-days to hatch by up to 13%.
Although
an
exposwe-response
trendwas evident, the two siqnificant
responses
were
not
consecutive.
Weights, moisture and percent
mortality of larvae at hatch, and
between hatch and swim-up varied
among treatments .
However the
varia-tions with exposure were
erratic and not siqnificant
.
Percent deformities varied
minimally (range O to 2.59%) but
replicates violated the assurnption
of homogeneity of variance and no
statistical analyses were performed .
Yolk sac conversion
efficiencies also varied erratieally
Although changes were
siqnificant by ANOVA, none were
different £rom control by Dunnettts
test.
.
For fry, an exposure-related
increase in mortality was only 12%
greater than control and was not
considered siqnificant (Table 5).
However, mean weights of fry
decreased by up to 51% significantly following an exposureresponse relationship (Figure 3).
Reduced larval weights at
hatch were the only measured responses of eggs to CP. Weights
followed an exposure-response trend
with a 17% decrease from control
weights at the highest test
concentration.
After
hatch, mortality of
larvae increased by 25% over
control but only at the highest
exposure level (Table 6). At this
test concentration, larval weights
at 10 days post-hatch decreased by
29% relative to controls (Figure
41, following an exposure-response
relationship.
Weights of whole
fish at the highest test concentration also declined by 18%
relative to controls but differences in yolk weights were not
related to treatments.
At the highest test level,
percent mortality of fry (Table 6 )
increased abruptly by 27%) corresponding to an increase in mortality
per 1000 fish-days £rom 3.4 to
13.1.
Weight at 4 weeks post
swim-up was also reduced, by 38%
relative to csntrol.
Percent
moisture at 4 weeks post swim-up
showed siqnificant changes, but
none were different £rom control.
For eggs exposed to DB, mean
larval weight at hatch decreased
significantly,
following
an
exposure-response trend (Table 7) .
Weight reduction at the highest
test concentration was 27% £rom
control. Percent moisture at hatch
decreased significantly at the
hiqhest exposure level, by 3%
relative to control.
After hatch, percent mortality
of larvae exhibited a sharp
monotonic rise, increasing by 70%
at a mid-range test concentration
(Table 7). Mortality levelled off
at the final two test levels.
There was an exposure-related
decrease
in larval weiqhts at
post-hatch.
We iqht
10 days
reduction at the hiqhest test
concentration was 52% relative to
Althouqh yolk weiqht
control
varied significantly, only weiqht
at the lowest treatment vas
different £rom control. This was
considered coincidental.
.
There was a sharp monotonic
rise in percent mortality of fry
(Table 7 ) , which reached a plateau
at a mid-range test concentration
(Figure 5). Mortality increased by
as much as 83% over control,
csrrespondinq to an increase in
mortality per 1000 fish-days £rom
2.0 to 66.7.
Total number of
fish-days at the hiqhest test
concentration were reduced eompared
to control.
Fry weights at four weeks post
swim-up were different £rom control
at al1 test levels. Only three
levels were
tested as hiqh
mortality had depleted the number
of Ery at the two hiqhest concentrations of DB to less than
10 per tank. Weiqhts were reduced
by as much as 70% relative to
mean control values (Figure 6).
After exposure of eqqs to TCB,
the mean larval weiqhts at hatch
decreased relative to control
The F-ratio for the
(Table 8).
analysis of variance was 2.94,
slightly k l o w the critical value
However, Dunnett's test
of 3 , 3 3 .
indicated that a mean weiqht
reduction of 15% at the hiqhest
test concentration was siqnificant.
A yolk weight increase
relative to control of 8% at the
highest test concentration was
also siqnificant, but no differences were observed for weiqhts of
whole fish or percent moisture at
hatch
.
Larval mortality increased at
the hiqhest test concentration by
60% relative to the controls
(Table 8).
Percent deformities
varied siqnificantly £rom controls
but with no obvious dose-response
relationship; the
result was
considered coincidental.
Larval weiqhts at 10days
poat-hatch
showed siqnificant,
monotsnic reductions, culminatinq
in a 41% decrease relative to
control.
Yolk weiqhts increased
over controls for al1 test levels,
but only two responses were
siqnificant. Weiqhts of whole fish
decreased
monotonically
with
increasing test concentrations.
Al1 weight losses were siqnificant
and culminated in a 21% weight
reduction relative to controls. A
37% decrease £rom control of yolk
sac conversion efficiency was
observed at the highest exposure
Hean respünses o f rainbüw trout to p-nitrophenol exposure and results of analysis of variance and Dunnettls t e s t
3
p-nitrüphenol cor~centrationlyH)
1 Hortality nf eyed egg,;
Z Hatch o f eyed eggs
Z Died vhile hatihing
Degree days tü hatct~
Net weight of "laïvae ün1y"at batch
Dry reight of "larvae ünly" a t hatch
Net weight of yùlk saï a t hatch (mg)
Dry weight ü f yulk sac at hatch (mg)
l e t weight uf 1div.i~and ynlk sas a t
Dry weigt~tnf larvae and yolk sac a t
X Hùistui-e 3t hatih
(mg)
isg)
hatch (mg)
hatch (mg)
0. O
3.0
97.0
2.0
381.0
24.6
2.8
43.0
17.7
67.6
22.5
88.0
p-nitropllenul iüncentïation !pH)
0. 0
1 Hw-tality of laivae
7.0
Z Defoïrired
0.34
Degree days fcütt hatch to swim-up
Met weight of "larvae ünly" a t 10 days (mg)
47.4
l r y veight of "Iarvae onlys a t 10 dayç (mg)
7.6
Wet weight of yolk. sac a t 10 days (mg)
35.8
D i y weight of yolL .sa< a t 10 days (mg!
16.2
Net neiqht u f lareae and yülk sac a t 10 days (mg) 83.2
Dry weight of laril.ae .and yolk sac a t 10 days ( ~ t g ) 23.8
Yülk sac cortvei~iuri efficiency iwet)
3.26
Yülk saï conversiun effiiiency idry)
1.59
Z Hüi5ture at 10 dàys
84.0
1
1
3
2
3
5
Sd
F
3.88
1. 0
99.0
0.0
398.0
22.5
4.3
42.6
19.3
65.1
23.6
82.0
7.19
4.11
96.0
2.0
411.0
16.3
2.7
33.9
17.7
56.2
21.4
83.0
8.99
4.0
96.0
3. 0
431.01
19.51
2.6%
39.6
18.4
53.1
21.0
86.6
16.46
4.0
96.0
2.0
415.0
18.7
3.2
43.4
18.9
62.1
22.1
83.0
28.75
14.0
86.0
10.0
432.0%
18.1
3.8
42.5
18.9
60.6
22.7
79.0
8.40
8.40
6.40
17.30
5.50
1.20
3.60
1.60
6.70
1.40
4.00
6.85
6.85
5.22
14.12
4.49
0.98
2.34
1.30
5.47
1.14
3.26
1.01
1.01
0.36
3.96
1.37
0.99
0.81
0.72
1.02
1.82
1.60
2.73
6.0
0.34
5.97
7. il
0.43
7.19
12.0
1.48
13.80
6. 0
21.57
8.0
2.59
3.80
1.27
3.10
1.04
0.83
48.0
7. O
30.6
15.0
78.6
22.1
2.71
0.76
85.0
42.4
6.7
33.9
15.4
75.3
22.1
3.Z
1.83
84.0
38.1
6.0
33.7
15.3
71.7
21.3
1.90
0. 8'3
84.12
43.2
6.7
35. 0
16.4
78.1
42.9
6.4
34.0
16.4
75.9
22.9
1.83
0.63
85.0
13.80
2.50
5.40
2.00
10.60
1.60
0.99
0.37
1.40
11.26
2.04
4.41
1.63
8.64
1.30
0.81
0.30
1.14
0.22
0.14
0.33
0.24
0.34
0.64
1.41
3.71
0.27
O.il
.>
i .-s . 1
. ~
2.14
CI =jSj
05. O
.
D
X
diff.
44
0.33
0.60
59
1
TABLE 5 (continued)
Mean respünses
--- -. - - - -
düf
---
rainbüw trout to p-nitrophenol exposure and results of analysis of variance and Dunnett's test
-
---
- --
n
..................................................
3
3
3
3
2
3
s
Sd
F
D
x
diff.
p-nitrophenol concentration (pH)
Z Hortality of f id)
Hürtality per 1000 fish-days
Total nmber of fish-days
Met weight at wrek 4 püst swim-up (mg)
Dry weight at week 4 post suii-up (mg)
1 Hoisture at week 4 post swim-up
O. O
20.0
5.0
7281
452.3
87.4
81.0
2.08
23.0
6.0
8368
378.0
65.4
83.0
4.17
24.0
7.0
7850
366.3
64.6
83.0
4.82
27.0
8.0
6334
358.0
58.8
84.0
10.50
30.0
9.0
6260
322.7+
51.6
84.0
17.97
32.0
7.60
11.0
2.50
5196
1140
221.3*
36.60
35.5*
14.20
84.0
1.70
6.20
2.04
930
29.86
11.59
1.39
1.20 21.59
2.17
4.16
3.20 3403
3.90
92.88
4.41
36.04
1.64
22
48
20
41
5, standard deviatiün calculated froi pooled error variance; sd, standard difference betueen means used to calculate D value for Dunnett's test;
F, F value frie ANOVA; D, ainiaue significant difference from contrùl in original unit5 as calculated by Dunnettls test; X diff., iinim~j~n
~ignificantpercent dif ference from control.
*
II
Treatment means signifiiantly different f r o i control by Dunnett's tejt.
_ 8,
= CIO d i t a .
Y = -158.06 LOG X
+
452.00 ( n = 5 , r2= 0.82 1
R = Log segression threehold astirnate
81 = Log bootstrap threrhold estlmate
G = Log geemetrlc rnern threshold estlmate
C = Control responsa
8.4
0. O
LOG
Figure 3.
0.8
1.2
P- NITROPHENOL CONCENTRATION (,Y, M
The effect of exposure to p-nitrophenoî on the wet weight of
ralnbow trout fry after four weeks of feeding.
Threshold effect
concentrations are shown for each statistical method of estimation.
1.6
Y =
- 34.21
R
B
G
C
-0.0
+
72.08
(
n = 3, r2= 0. 9 9 )
Log regression threshold estimate
Log bootstrap threshoid estlmate
Log geometrlc mean threshold estimate
Control response
-0.4
LOG
Figure 4.
=
=
=
=
LOG X
8
1
II
I
\
R
0.4
8.8
P-CHLOROPHENOL CONCENTRATION (p M 1
The effect of exposure to p-chlorophenol on the wet weight of
"larvae only", 10 days after hatch.
Threshold effect concentrations are s h o w for each statistical method of estimation. The
regression is based on the filled circles.
Hean respunses of raintiüw t r o u t t ü p-chlürophenol expüsure and r e s u l t s of a n a l y s i s of variance and Dunnett's t e s t
p-chlorophenol concentration (pH)
Z H o r t a l i t y ü f eyed eggs
Z Hatih of eyed eggi
1 Died while hstihing
Degree days t o hatch
Net weigt~t o f " l a r ~ a eu n 1 y " a t hatih
Dry n ~ i q t i t ü f " l a r v i e ünly" a t hatch
Net weight of yoll.. sac a t hatch ( i g )
Dry weight uf ÿülk sac a t tiatih (mg)
Net weight o f larvae and yùlk sac a t
Dry weight of larvde and yülk sac a t
Z Hüisture a t hatrh
(ig!
iig)
hatch (mg)
hatch (mg!
0.0
0.0
100.0
0.0
364.8
30.6
4.2
50.8
23.6
81.4
27.8
86.2
0.24
0.49
3.13
6.5
97.0
93.5
0.6
1.6
358.9
369.5
31.3
16.3*
3.7~ 3.5~
48.6
43.8
22.3
22.6
79.9
76.1
26.6
26.1
88.3
86.9
LARVAE
p-chlürüphenül cuncentration ( l n )
Z f l u r t a l i t y i f larvae
Z Deformed
Degree days frûin hatih tu ;~iin-lip
Net weigtit ü f "lat-vae unly" a t 10 days (mg)
Dry weigt~tof U l a r v a e unly" a t 10 days (mg)
Net weight of yolk sac a t 10 days (mg)
Diy weight u f y ~ l ks a ï a t 10 days ( i g )
Net weight of larvae dnd yolk sac a t 10 days (ug!
Dry weight of larvae and yulk sac a t 10 days (pql
r i- l k sac conversiun e f f i i i e n c y (wet)
Yolk j . 3 ~i ~ n ~ e ï s ~ eùfff i i i e n c y ( d r y ?
Z Hüisture a t 10 d i p
11
0.0
0.24
4.0
6.1
13.2
15.1
99.2
99.6
72.6
74.3
9.9
1i3.1
32.0
27.7
15.9
13.8
105.6
102.0
25.8
23.3
i .78~ 2.13
7
0.73
0.71
86.6
06.3
0.97
5.2
94.8
2.4
365.7
27.8*
3.6r
50.7
23.6
78.5
27.3
87.1
1.94
2.3
97.2
1.2
361.3
27.7*
3.64
51.1
23.3
78.0
26.9
87.0
3.85
1.4
38.6
0.a
Z67.7
25.3*
3.5*
50.0
22.9
75.3
26.4
86.2
3.88
3.88
1.38
4.36
0.60
0.14
1.77
0.93
2.60
0.77
0.78
3.54
3.54
1.26
3.98
0.54
0.13
1.61
0.85
2.37
0.70
0.71
2.47
48.80
7.90
0.77
0.56
2.07
1.69
2.67
1.72
0.41
6
10
TABLE
b
(continuedi
Mean responses of rainbow trout to p-chloropbenol exposure and results of analysis of variance and Dunnettls test
~
-----
--
--------
-
~
-
n
...................................................
3
3
3
3
2
0.97
20.3
5.2
8814
496.6
98.1
80.4
1.94
17.9
4.7
8568
480.7
04.6
80.3
3
5
s
d
F
O
%
diff.
FRY
p-chlorophenol concentration ($4)
% Hortality of fish
Hvrtality per 1000 fish-days
Tt:ltal nunber üi fish-days
Net weight at week 4 post swim-up (mg)
Dry weight at week 4 post swie-up (mg)
Z Hüisture at week 4 püst swim-up
O. O
15.2
3.4
9653
491.1
95.4
80.6
0.24
0.49
16.0
20.5
5.0
3.9
3157
8642
470.5
512.2
33.2
101.6
80.2
80.2
3.85
42.6s
9.06
8.27
13.lt
3.10
2.83
12?2
L098
7613
304.7s 61.00 58.43
56.8% 13.00 12.14
81.7
0.55
0.50
3.73 26.30
3.90
9.00
O. 84
4.17 185.81
4.56 38.60
1.59
3.63
26
38
40
2
5, standard deviatiün calculated froa pooled error variance; sdl standard difference between means used to calculate D value for Dunnettls test;
FI F value fïü~nANDVA; Dl minimum significant difference from cüntrol in original unit; as calcultated by Dunnettls test; X diff., minimum
significant percerit difference froi control.
*
Tieatrent mran; ,3ignificantly different fror control by Dunnett's test.
Hean responses of rainbow t r o u t t u 2,4-dichlorophenol exposure and r e s u l t s of a n a l y s i s of variance and Dunnettls t e s t
3
3
3
3
2
3
S
S
d
F
D
Z
diff.
2,4-dict1loropheno1 cüncentration (pH)
1 Hovtality o f eyed eggs
Z Hatch of eyed eggj
X Died while hjtihiiig
Degree days t a hatck
Net weight o f " l a r v i e u n l y h t hatch (eg!
Dry weight o f "larvae o n 1 y " a t hatch (sg)
Wet ueight
folk 5ai a t hatch (mg!
Dry weight of ~ ü l tsac a t hatch (agi
Net weighk
larvae and y ~ l ksac a t hatch (mg)
Dfy weight.of larvae and yolk sac a t hdtch (mg)
Z Haisture d t h a t i t ~
lof
lof
- 4-dichloropt~enc~lconcentration (pH)
0.0
0.6
99.4
0.0
361.3
27.3
4.2
48.0
23.0
75.2
27.3
84.8
0. 0
3.7
0.2
Wet weigt~t1 j f "lat-vae u n 1 y " a t 10 days (mg)
57.8
Dry weigtlt (if "larvae only" a t 10 days (mg)
15.4
Met weight of yuIl; sac a t 10 days (mg1
44.9
Dry weight uf p l l : s a ï a t 10 days (mg)
16.5
Wet u i i q h t of larvae and yolk sac a t 10 days (ag) 107.2
Dry weigt~t of larvae and yülk sac a t 10 days h g ) 31.9
Yülk sac ioiiveiji~jne f f i i i e n i y Iwet)
0. 90
Yolk sac cunversion e f f i c i e n c y i d f y j
1.75
2 Hoijtuie 35 1(:! days
75.0
?I
Z Murtality of larvae
Z Deforffied
0.61
0.3
99.7
0.13
364.5
24.7
3.8
43.4
23.6
24.1
27.4
84.7
1.10
il.9
39.1
0.17
372.8
24.6
3.9
49.8
27.2
74.4
27.E
84.1
1.96
1.1
98.9
0.46
375.6
21.5,
3.6
51.8
24.8
72.7
28.4
83. O
TABLE 7 (cuntinued:]
Hean responses o f rainbow trout to ?,4-dichlorophenolerposure and results of analysis of variance and Dunnettfstest
3
3
3
3
2
3
s
sd
F
D
Z
diff.
FRY
2,4-dichlorophenol concentration (pli)
Z flortality of fish
Hortality per 1000 fish-days
Total number o f fish-days
#et weigt~t at ueek d post swie-up h g )
Dry weiyht at week 4 pnst swie-up (mg)
Z Huisture at week 4 püst swia-up
0.0
0.61
1.10
1.96
6.6
8.8
43.8*
86.2,
2.0
2.9
16.2
47.0t
6358
8029
6625
3974
704.0 427.3, 263.3* 212.0t
3.44
88.2*
54.D
3863
6.14
89.7, 7.08
5.78 31.40
18.38
66.7t
7.44
6.07 43.07
19.30
3667t 1230
1004
7.20 3191
48.58
39.50 44.80 125.61
18
46
18
S, standard deviation calculated from pooled error variance; sd, standard difference betueen ieans used to calculate D value for Dunnettlstest;
F, F valu^ ft-üü~ANOVA; D, mi~ieuesignificant difference from contrül in original units as calculated by Dunnettls test; 1 diff., minimum
siynificant peiiint difference from cinkrol.
r Ireatmnt aedr~ssignificantly different frum contrùls by Dunnettrstest.
'-' = no data.
concentration, but there were no
changes in percent moistue.
Fry mortality increased by 62%
over controls at the highest test
concentration, corresponding to an
increase in mortality per 1000
fish-days
of
3.71 to 43.4
(Table 8).
No significant changes
were recorded for the total neunber
of fiski-dayç.
Fry
we ights
decreased
significantly, but tests were
performed on only four treatments,
as the number of fry in each tank
had dropped to less than 10 at the
highest test concentration. Weight
changes
were
monotonic
and
culminated in a 53% reductisn
relative to control (Figure 7).
Slight
increases
in percent
moisture for fry of about 2 to 3%
relative to control were significant, but responses were not
monotonic.
Results for P B were obtained
from Hodson and Blunt (1981).
Although response means could not
be
compared statistically, the
occurrence of exposure effects, as
determined in the original publication by ANOVA (p 5 0.01), are s h o w
in Table 9.
P B toxicity to eggs are not
discussed here and the reader is
referred to the original work.
The ANOVA indicated significant
increases
in
larval
mortality per 1000 fish-days. At
the highest test concentration,
this corresponded to a change £rom
contsol of 11.84 (Table 9 ) .
We
assmed that this response was
significant as al1 other test
responses were similar to control.
Weiqht of larvae at swim-up
decreased
monotonically
and
reduction
culminated in a 15%
relative to control.
Yolk sac
conversion efficiency decreased
relative to control by 14% at the
hiqhest test concentration.
For fry, mortality per 1000
fish-days increased by up to 72.0
at the highest exposure level
Only this value was
(Table 9 1 .
different £rom the control value of
A siqnificant concentration
4.0.
effect was observed for weights of
fry at 4 weeks post swim-up.
Weights decreased monotonically,
culminating in a 72% decrease
relative to the mean control value
(Fiqure 8)
.
EFFECT THRESHOLD ESTIMATES
Reqression estimation
A significant exposure effect
was observed for larval weight at
hatch. Weights decreased monotonically with a 27% reduction relative
to control at the highest test
concentration. Several symptoms of
Mortality
parameters were
less
amenable
to regression
analysis than weight parameters, as
mortality often increased only at
the
highest
exposure
level
Y = 153.08 LOG X + 39.97 ( n = 3 , r * = 0 . 9 9 )
R = Log ragresslon threshold estlmate
B 4 Log bootstrap threshold estimrte
G = Log geometrlc m a i n threshold estlmate
C = Controt response
LOG
2,4
- DICHLORBPHENOL CONCENTRATION (PM
)
Figure 5. .The effect of exposure to 2,4-dichlorophenol on the total mortality
of rainbow trout four weeks after swim-up.
Threshold effect
concentrations
are
shown
for
each
statistical method of
estimation. The regression is based on the filled circles.
ra
E'
3
-- C
Y = -426.80 LOG X + 318.42 ( n = 3 , r 2 = 0.92 1
w
R = Log regression threshold estimate
B = Log bootstrap threshold estimate
C = Csntrol response
600-'
I
Z
zcn
c..
cn
P
d
g
400-
W
$!
I
I
I
1
I
I
I
5
I
I
C-
I
I
I
I
I
X
O
I
W
200-
I
III
C
E
B
O
- 0.4
I
LOG
Figure 6 .
1
1
t
0
2,4 -DICHLORBPHENOL
1
0.4
l
f
0.8
( pM 1
The e f f e c t of exposure t o 2,4-dichlorophenol on t h e wet weiqht of
fry four weeks a f t e r swim-up. Threshold e f f e c t c o n c e n t r a t i o n s a r e
shown f o r
each statistical method of e s t i m a t i o n except t h e
qeometr ic mean.
(Table 9).
The results of the
regression
analyses
and the
interpolated 25% minimum change
thresholds are given in Table 10
for those prameters that met the
criteria listed in the Materials
The correlation
and Methods.
coefficients (r) in al1 cases were
greater than 0.85, indicating that
the regressions accounted for more
than 72%
(rz>0.721 of the
observed variation in responses.
Except for TCB, slopes (b) for
the regressions were similar among
experiments for wet weights of
larvae at hatch and for wet weight
at 10 days post-hatch (Table 10).
Greater differences among experiments were observed for the slopes
of wet weight of fry at 4 weeks
post swim-up and for mortality
parameters.
N O K - L O K seometric mean
The LOEC values for those
parameters which were significantly
different £rom control, a d which
differed by 25% or more, are listed
in Table 11, along with the
carresponding NOECs and geometric
means. Reductions in larval and
fxy weight, and increases in
mortality were the most consistent
responses. While m n y changes were
observed in percent moisture, the
extent of changes were uniformly
small and did not meet the
criterion of a 25% change.
Al1 chemicals
except NP
elicited
a
significant, 25%
minimum reduction in weiqht for at
least one of the two larval weight
parameters (weight at hatch and at
10 days post-hatch) (Table 11).
For NP, larval weight at hatch was
reduced 37% relative to controls
(Table 51, but this reduction was
not significant due to large
within-group variations.
A 57%
needed to
reduction would be
demonstrate a positive Dunnett's
test.
Al1 chemicals except PH caused
a significant 25% minimum reduction
in fry weight. Although fry weight
declined by 27% in the PH test
(Table 4), this change was not
significant and a minimum percent
difference £rom control of 70%
would t>e required for a positive
Dunnettfs test.
For other
chemicals, the percent difference
£rom control needed to demonstrate
a significant change by Dunnettfs
test was 18% for larval weight and
20% for fry weight.
In the PCP
experiment, both larval weight at
hatch and fry weight at 4 weeks
post swim-up decreased relative to
control by more t b n 25%. Significant
treatment
effects were
demonstrated by ANOVA at a more
conservative probability level of
0.01 (Table 9). Thus it is likely
that the 25% minimum weight
reductions observed for the PCP
experiment represent significant
reductions from control values.
LOECs based
on dry weight
measurements are also listed in
Table 11 but are the same as those
based on correspondinq wet weights.
Weiqhts of whole fish demonstrated a siqnificant, 25% minimum
reduction for the PH experiment
--
sr,
-.
-
p.,
7
,
'
- ,:.,
p.,
'a
,.a
,:.,
d
-
p,
m-r ~ ~ n m h c n , = ~ r - . i ~ r c u ~ ?
r-. --<
m u-i t r r r? r5.i m sm n
C'.A ~1r.4 m
-. 3 5 r-. e.4 r 3 s
a r.3
-r m r . 4
r.1
W
. . . . . . . . . . .
-
-. rr.
. . . ,-'.,
. . . . -. . . '=*
C.
8
9 .
a m rm tm 8
s
m m c-4 -r a r - a r-ta
,y
03
O d a p., ,=) C.
U7
-
*
*
=
+
a
*
*
-
m
5,
*
X
-%Yu3
. l. m. u-,. -. UT. a. . . . -. .=,. .
w u
TT
.-1
P
)i
h 2
0
.
t
g
C.4
CD 'Y Ln C r i
-
O
C 1
t P
U?
iJi
UT8
? 8
2
3
7 i7'i
PLI
?3.
?.4
CU
. . r-. . m. . u-, . w. -. . -. m. .
LI?
P.
CQ
C
'
.
,
CO
--i
--
m
c.3
p.4
id, 0 'Y C 1
117 *?A
-r
m
Lm
ici3 P.4 lu,
b
u
C'.r
m
. . r-.. . . . . . r. . . a. -.
--.
m
c.3
C'I h C'.a
in
%Yrf
C
r
c..r
ri-,
U 7 'd.2
V3
Cri
7S' .ce CO
rf
p.,
-%Y
rf ,.a:
C.
>?.,
h
m
n-3
- - z - - * *
si(
u m m m
u *
<- E
in m .,ou
I=
-
.-..
.a
:
1
-
u
.-.
7III
an m
Ln L n
+a .+3
3 4
8
2
am , n . , A . b . x x
=
.CC
x
2
w
c
in
r-,
,.n
C I 2
5-
iu
23,
73
T;TI
;i l 3
cri
,II
G
;
-(1.1
ln -,
.CI
L
,>
su.,
.4-
L.8
.
P
.
,
L
il,
:
.*
8n
---,
,:
C
<=,
G
,n
A
.
#CS
Ir,
3
-n
IU
.<
IL-
,TI
'Tl
A
x
TL
c2,
!, T h
=-<:
r.e
N
a
lu
..-.
Y--
,+
U
L
.
Ti
w
3
nz,
ixi
!C>
L
Cri
-4 -3
8x1
TL
-.n o .y;-
-3
2".
CI
2
.
-
.1b.3
.
.,-.
m ^ n o
C
u
in
a.n
d
m
,U
m
>
:-
,=, ,ci ,n
=
Y-
-$1,
-.a
z.3,
=
.&Y
iu
e
=.m
'-
?La
4n
77
iu
>:
iT1
A=
i3.I
iu
cc, c
c
id
m
,O
L
.=
C1
.3
4
'
L,
IO
L
A
cs,
-L~I
i
i
.
*-
ILI
3.
8
..-.
A=
.c
-.=
.
1'51
c c
,=, ici
y:".
.+
,z,
C
d
m
n
31,
!-
,..-,
-3
-Y
c G c
u i
CI, m
3
'+-
81.
--
iu
+i
3
3
-*--i
+' <'.
<nt
-3
:-
n
a
*U
2;
L
~.-e
.
L
.<
.3
!,-
d
4
.,-
i
e
n
Li
72
il,
,E
--
II,
C.
.1
72,
L
ici
- T ' L +
1
-.
:.
,
.'7.
4.2
a=
,=,
11.1
'-8
u
ri'i
Ici
1x
a
.
3
-4 -4 L
-
7-3
iu
U
Il,
:-
.I,
:,.+
,
,,
..-
u
",
:-
:=
L
118 ui
3
in
,U
m
a
.+ .-'U
3
C
8n
=>
2
,-
$2, ,3
Z
.h'h
u
1
.-ni
8
m m - -
=
Uf
E=
a
*.-1
2-I
il,
i=
Cn
.+
il,
I
I
i 3
L
CT, -c;.
il, 111
O
3
42.
.c i
:C
CI, CTi ni
U
3
m
u
2
.c<
i l ,
.3
.-I
0,
-3
i-1
I-i
n
111
:n
u
,m
$1,
,u
-.-
>i(
3
,.ri
4
-iL
?...
ici
E:
3-
i-e
3.8
Mean i e i p u r ----i ~ e sof ï a i n b o ~t i ü u t t u 1 , 2 , 4 - t r i i h l ü r ü t j e n z e n ~expùsuïr 2nd r e u l t s of a n a l y s i s
J
l I 2 , 4 - t r i ~ h l l i l i o h e n z i . ~ i ionieiitratiün
e
(pH)
1 McS:t.jlity o f fi.:h
Murtalitÿ pei iiji10 fiçh-day;
Tut i l nui~beiv t fish-days
Net weiqt~t 3 t ~ f e e k4 p ~ s tsuiri-up img,
0 ï y ueiqht .jt weel 3 pe,jt suim-up (~ng)
1 Ruistuïe a t wrek 4 p s t swim-up
3
0.0
0.55
13.1
7
3.71
5.0
4375
3536
539.0
586.0
110.0
120.0
74.5
73.b
9
i
O.8E
23.2
'3.0
3382
420.0%
30.0
8l.it
irif
~ v a r i a n cand
~ Dilnnettls t e s t
3
.?
1.33
23.1
6.7
3320
408.0n
78.0
81.0
2,jg
27.7
7.8
4055
252.gt
44,c:it
at.7t
*J
S
5,
9-42
74.6t
'3.17
7.48
45.4%
4.90
4.00
2031
1405
1146
7.50
6.12
16.2
12-22
0.61
0.50
F
1b.70
28.53
D
h
diff.
23.26
12.44
23
19.16
41.38
1.56
4
38
2
0.99
3.05
10.21
13.65
5! s t a n d ~ i i l d e i i i i t i n n c a i i i l a t e d f r ü r pvoled errui- sariarrce; s, standsrd di?feuence tletween iriearis used to calctilate D u a l i ~ ef o r Dunnettls t e s t ;
F, F .:iii,r- + r i i i i A. i i ü ~ ~D,; irilnielum s i g n i f i c a n t d i f f e r e n i e f ï v r i o n t r a l i n n ï i g i n a l units; X d i f f . , imlnimuili s1griificant p e ï i e n t d i f f e t e n i e fi-uin
itjntïoi.
'
-O,
Tieatiiier!t ,!i;aris iiignifiiiiitly d i f f e r e n t fiom cüntïvl DY Duririettls t e s t .
Y = -400.70 LOG X + 454.95
R
B
G
C
O r
-0.4
l
0
LOG
Figure 7.
l
(
n = 4, r2= 8.81
= Log regresslon thseshoid estlmate
= Log bootstrap threshold estlmate
= Log geometric mersn threshold estimate
= Control response
I
0.4
I
I
1
0.8
1,2,4, -TRICXLBROEIENZENE CONCENTRATION
(pM )
The effeet of exposure to 1,2,4-trichlorobenzene on wet weiqht of
rainbow trout four weeks
after swim-up.
Threshold effect
concentrations are shown for each method of estimation.
TABLE 9
Mean responses of rainbow trout to pentachlorophenol exposure and results of analysis of variance
"
"
n
7
pentachlorophenol concentration (WB
Mortality of eyed eqgs
Degree days to hatch
Wet weight of "larvae only" at héch (mg)
%
2
2
O. O0
1.O
305.00
O. 0 4 1
0.6
323.00
18.5
19.1
- .
.
"
.- ........ ....
-
-
2
2
0.105
O. 3
336.00
15.8
CONCENTRAT1ON'
EFF'ECI"
O. 308
1.7
341.00
14. O*
pentachlorophenol concentration (pl)
Mortality per 1000 fish-days
Wet weight of "larvae only" at swim-up (mg)
Yolk sac conversion efficiency (wet)
YES
YES
E S
FRY
pentachlorophenol concentration (pM)
Mortality per 1000 fish-days
Wet weight at week 4 post swim-up (mg)
*
-
0.00
4 .O0
847.00
O. 060
2.96
728.00
O. 094
2.94
628.00"
O. 030
72.00
234. OO*
YES
llES
Treatment means differing from controls by 25% or more.
Original factorial analysis of variance tested exposure, temperature, and exposure-temperature interaction effects
at the 99% confidence level (Hodson and Blunt 1981).
only (Table 11). The corresponding
LOEC value was about 4 times higher
than the lowest LOEC for that
experiment, which was based on
reduction of larval weight at
hatch.
Yolk
we ight
conversion
efficiency
demonstrated
a
significant, 25% minimum reduction
for the TCB experiment only
The LOEX value vas
(Table 11) .
about 1.5 times that of the lowest
LOEC for that chemical based on
reduction of fry weight.
one test concentration level (Table
11); no LOEC waa based on mortality
F r 1000 fish-days in the CP test.
L0EC.s for hatch parameters were the
same as those for mortality in the
TCB expriment.
No mortality
mrameters were significant in the
NP study.
In the P B test,
mortality per 1000 fish-days was
the
only mortality parameter
available. From Table 9 it can be
seen that fry mortality increases
at the highest test concentration
of P B , with no change at lower
exposure levels. Thus it is safe
to conclude that the LOEC value for
this parameter response was the
highest PCP concentration.
Al1 chemicals, except NP,
elicited
a
siqnificant, 25%
minimum increase in mortality
(Table 11).
For NP, percent
mortality increased by only 12%
relative to control mortality
(Table 5), and a 19% increase
would be required for a positive
Bunnett's test.
The only other parameters to
m e t the criteria were those of
yolk conversion efficiencies of
TCB
The NOELs and LOECs were
similar to those associated with
weight changes (Table 11).
Mortality per 1000 fish-days
is cumulative, and is based on the
nwntser of fish in a tank, which
decreases with time.
Computer
simulations were performed to
estimate the increase in mortality
per 1000 fish-days necessary to
increase the percent mortality of
fish by 25% over the period of fry
These
simulations
growth .
indicated that a 10-fold increase
relative to control would be
sufficient.
The lowest geometric mean
estimates for reduction of larval
or fry weights and for increased
mortality are listed in Table 12.
Weight reduction thresholds are
lower than those for mortality
except in the case of CP where
theyare the same. The geometric
mean for weight reduction in the
iXP experiment is given as a range
since the NOEC was O uM (control).
These threshold estimates are
cruder than those estimated by
reqression analysis.
LOECs
for
percent
The
mortality of fish in the PH and
DCP experiments were lower than
those
for mortality per 1000
fish-days by a factor of 1.8, or
A comparison of the regression
estimates of thresholds with the
corresponding
geometric
means
(Table 11) indicates a similar
trend and magnitude in values. As
.
I
- 1.2
1
f
1
-.
I
I
-.8
4
LOG PENTACHLOROPHENOL CONCENTRATION (FM)
Figure 8 .
1
O
The e f f e c t of exposure t o pentachlorophenol on t h e wet weight of
rainbow t r o u t a t four weeks a f t e r swim-up.
The threshold e f f e c t
concentration was e s t i m t e d £rom t h e r e g r e s s i o n .
TABLE 10
Linear r e g r e s s i o n s r e l a t i n g measurements t o logarithm of concentration ( y = a + b l o g XI.
Thresholds a r e i n t e r p o l a t e d values a s s o c i a t e d with a 25% chanqe r e l a t i v e t o c o n t r o l measurements.
Interpolated
threshold
b
phenol
Wet weight of "larvae only" a t hatch
Dry weight of "larvae onlyv a t hatch
Wet weight of "larvae onlyff a t 10 days
Dry weight of "larvae only" a t 10 days
Wet weight of l a r v a e and yolk s a c at 10 days
% mortality f ish
M o r t a l i t y per 1000 f i s h days
Wet weight a t week 4 post swim-up
Dry weight a t week 4 post swim-up
Wet weight of "larvae only" a t 10 days
Wet- weight a t week 4 post swim-up
Dry weight a t week 4 post swim-up
% m o r t a l i t y larvae
Wet weight of lllarvae only" at hatch
Wet weight a t week 4 post swim-up
Wet weight a t week 4 post swim-up
% m o r t a l i t y larvae
% mortality f i s h
M o r t a l i t y per 1000 fish-days
r
(uM)
Linear regressions relating measurements to logarithm of concentration (y = a t b log X).
Thresholds are interpolated values associated with a 25% chancre relative to control measurements.
b
r
Interpolated
threshold
(uM)
Wet weight of "larvae only" at 10 days
Dry weight of "larvae only" at 10 days
Wet weight at week 4 post swim-up
Dry weight at week 4 post swim-up
Yolk sac converison efficiency (wet)
Yolk sac conversion efficiency (dry)
pentachlorophenol
Wet weight of "larvae only" at hatch
Wet weight at week 4 post swim-ug
n, number of observations; a, constant; b, regression coefficient or slope; r, correlation coefficient.
Threshold concentrations associated uith a 25% chanse relative ta contvol values.
M e t weight of 'larvae only' at hatch
*Dry veight o f 'larvae only' at hatch
tWet weight of "larvae only" at 10 days
*Dry veight of 'larvae only' at 10 days
tWet weight of larvae and yolk sac at 10 days
t X Hurtality fish
inortality per 1000 fish-days
p-nitrophenol
*Met weight at week 4 post swim-up
Dry weight at werk 4 post suia-up
p-chlorophenol
*Net veight of "larvae on1y"at 10 days
Yet veight at week 4 pùst swir-up
Dry veight veight at week 4 post suis-up
iI Hortality larvae
1 Hortality fish
*Met weight of "larvae only' at hatch
Wet weight a t week 1 püst svim-up
Met weight a t week 4 post swir-up
X flortality larvae
tl Hortality fis11
Hortality per 1000 fish-days
NOEC
LOEC
(uni
<un)
8.46
8.46
29.50
29.50
29.50
13.60
29.40
Geùmetric
aean
(un)
Bootstrap estimate with
85% confidence interval
(un)
Regression
estiiate
(un,
TABLE 11 (continued)
Threshold concentrations associated with a 251 chanae relative to control values.
NOEC
LOEC
(uni
(ufli
Geometric
mean
(un)
Bootstrap estimate with
951 confidence interval
Regressiün
estiiate
(un)
i UN)
N e t weight of "larvae only" at 10 days
%Dry weight o f "larvae only' at 10 days
*Met weight at week 4 post suia-up
Dry weight at week 4 post suim-up
Yolk sac cünverison efficiency (uet)
Yolk sai ianversiun efficiency (dry)
*T Hortality larvae
*1 Hortality fish
*Z flortality per 1000 fish-days
ttpentachlorophenol
*#et weight of "larvae onlyu at hatch
ilet weight at week 4 post swii-up
tHürtality per 1000 fish days ( f r y )
t
Responses exhibiting a 251 change relative to control which is statistically significant at the 952 confidence level.
*x
Dunnett's \pst nùt applicable to pentachlùrophenol data. See text.
Y-U
= insuffiiient data for Büotstrap estimation.
in the case of thresholds estimated
by the geometric rnean, the lowest
threshold estimated by regression
is based
on growth and not
mortality. Linear regressions of
weights vç log concentration are
illwtrated
in
Figures 2-8.
aairesholds
estimated
by the
regression method, geometric means,
and the bootstrap method are çhown
for comparison.
Bootstra~estimation
Bootstrap estimates of the 25
percent effect concentration were
calculated for those responses
which exhibited a significant
reduction relative to control
values by 25% or greater. Percent
mortal ity was
transformed to
percent survival (i.e. 100% - %
mortality) but mortality per 1000
fish-da*
was not amenable to the
Mortality
bootstrap procedure.
data for DB, and al1 data for P B
were mean values, and "bootstrapping" could not be performed.
Bootstrap threshold estimates
and their 95% confidence intervals
were quite similar to their
respective regression estimaites
(Table 11) This is understandable
given the good correlation coefficients (r) for the reqressions
(Table 10) and the fact that
bootstrap estimates are interpolated £rom an assumed linear
function between any two adjacent
means.
.
Regression estimates of the 25
percent effect threshold concentration for larval or fry weight
reduction were available for al1
compounds tested. These thresholds
were consistently lower (Table 11)
than those derived £rom mortality
Threshold concenresponses
trations for NP, DB, TCB and PCP
were bsed
on a
25% weight
reduction for fry at 4 weeks post
The CP threshold was
swim-up.
bsed on a 25% reduction in larval
weight at 10 days pst-hatch but
the threshold associated with
reduced fry weight was essentially
the same as that for larval weight
(3.27 vs 3.45 uM).
.
The threshold concentration
for PH was based on a 25% decrease
in larval weight at hatch. While
fry weight was reduced by 27%,
response variability was such that
no significance was attached to
thia change. However, based on a
25% minimum change, without a
statistical significance criterion,
the N O K and LOEC for fry weight
reduction is 13.60 uM and 29.40 uM
(Table 41,
with a calculated
geometric mean of 20.00 uM. This
value is quite similar to the
regression estimate of 15.34 uM,
bsed on reduction of larval weight
at hatch.
Therefore, threshold concentrations associated with weight
changes of fry provide a consistent
measwe of chronic toxicity for the
development of QSARs.
QSAR
The final chronic threshold
concentrations
for
the
six
compounds are listed with 96 hour
LC5Os and the ratios of chronic to
acute toxicity (Table 13).
The
rank order of chronic toxicity was
similar to that for acute toxicity,
except for DCP, which had a qreater
chronic toxicity than CP or TCB.
The average value for the ratio of
chronic to acute toxicity was 0.13,
with a standard deviation of 0.066,
or 50% of the average value. The
ratio of chronic to acute toxicity
for PH, TCB and PCP were consistent, but more variation was
observed among the NP, 8 and DB
ratios; they deviated £rom the
average by about 30, 60 and 80%
respectively.
QSARs for acute and chronic
toxicity are given in Table 14 and
parallel
(Figure 9).
appear
Regressions
(F-test) and slopes
(t-test) for both QSARs were
significant at the 0.05 probability
level. The QSAR for acute toxicity
accounted for more of the variability amonq LC50fs (89%) than the
chronic mode1 for variability among
thresholds ( 75% 1
.
A t-test was performed to
determine if the slopes of the two
QSARs were homogenous. However,
poolinq of the sum of squares in a
t-test for small sample sizes
(n 5 30) is only legitimate if the
variance about the regression line
is the same for each sample.
Variance was assessed by calculating the ratio of the two variances
(F-ratio) and determining if they
were
statistically
different
(F-test) at the 0.05 probability
level.
No significant difference
was observed and the results of the
t-test (t = 2.67, deqrees of freedom = 81, indicated that the slopes
for the acute and chronic QSARç
were homoqenous p 5 0.05.
DISCUSSION
This experiment demonstrated
that the chronic toxicity of a
series of substituted phenols and
benzenes increased with increasing
octanol-water partition coefficients. Toxicity increased in the
order phenol < p-nitrophenol <
p-chlorophenol < 2,4-dichlorophenol
<
1,2,4-tr ichlorobenzene
<
pentachlorophenol.
The cornparison of control
responses amonq the chronic tests
provided perspective on treatment
to vnormlv
effects relative
variability.
Mean
control
responses for weights of larvae and
3)
demonstrated
fry
(Table
increasinq levels of variability
for successive life stages. The
standard deviations, calculated as
a percent of the overall means,
were :
8% for larval weiqht at
hatch, 17% for larval weight at 10
days post-hatch and 28% for fry
weight at 4 weeks post swim-up.
The
increased var iabi1i ty £rom
hatch to lO-days post-hatch may
indicate
di£ferences in yolk
utilization, as larval weight
increased by an average of 130%
during this period.
TABLE 12
Cornparison of threshold estimates based on larval or fry veiqht reduction and mortality increase.
COMPOUND
(NOEC-LOEC)
GEOMETRIC MEAN (UM)
WEIGHT
MORTAL ITY
BOOTSTRAP
WEIGHT
phenol
11.63
14.58
20.00
ESTIMATE (UM)
MORTAL 1TY
25.10
REGRESSION EÇTIMATE (UM_L
WEIGHT
HORTALITY
15.34
P-nitrophenol
7.11
7.81
5.20
P-chlorophenol
2.74
3.21
3.27
pentachlorophenol
21.33
Although
yolk conversion
efficiencies varied considerably,
with a 20% standard deviation, they
were not related to the percent
increase
in
larval weights.
Indeed, the highest control value
for yolk conversion efficiency was
associated with the lowest percent
increase in larval weight. The
standard deviation for weight of
yolk at hatch was 9%, and at
10-days post-hatch, 33%; the increased error of yolk sac weights
at 10-days post-hatch may be a
function of their smaller size,
since they were
less easily
dissected.
Conversely, larger
variation in the weights of larvae
at lO-days post-hatch is likely the
result of differences in yolk
utilization, since larger larvae
were more readily dissected and
could be weighed with less relative
error
.
Fry weight exhibited the
highest relative variation for
measured growth in this study,
which is characteristic for this
stage of development.. The fish are
no longer self-sufficient, and
differences in rations and aggressive behaviour of individuals due
to density-dependent cornpetition
for food and space will emphasize
differences in growth rates among
individuals
(Eaton and Farley,
1974)
Variations in experimental
conditions might also cause growth
differences
among experiments.
However, post-hoc comparisons of
growth responses were made within
experiments, where conditions were
more consistent.
.
Due to variability in growth,
an average weight reduction for
larvae and fry of 20% relative to
control was required for a positive
Dunnettfs test ( p 5 0.05). Since a
25
percent
weight reduction
relative to control was chosen as
the chronic toxicity end point, the
threshold for toxicity was outside
the bounds of normal variability.
There are many potential end
points for estimating the chronic
toxicity of chemicals to fish. In
this study, mortality and simple
growth, as measured by the weight
of larvae or fry, proved to be the
most consistent. Five of the six
chemicals tested caused significant
increases in mortality relative to
control, and al1 of the chemicals
decreased the weights of larvae and
fry.
Threshold concentrations
based on reduced larval or fry
weight were lower than those based
on a 25% increase in mortality by
an average factor of 0.48 (range:
0.25 to 0.67; calculated £rom
Table 121
.
For QSARç, weight reduction
was chosen as the end point to
ensure a complete and comparable
set of chronic toxicity threshold
concentrations.
However, only four of the
final six chronic toxicity end
points could be based on a single
end-point:
fry weight reduction.
A 25% reduction in fry weight was
not statistically significant in
the case of PH and 8 due to considerable within-group variation.
Therefore,
we
chose
weight
reduction during the larval stage
as an appropriate substitute. End
points for PH and 8 were based on
LOG T'
I
0
I
I
I
I
6
2
LOG P
Figure 9. The relationship between sctanol-water partition coefficient and
measures of acute and chronic toxicity for phenol, p-nitrophenol,
p-chlorophenol,
2,4-dichlo~ophenol, 1,2,4-trichlorobenzene and
pentachlorophenol.
TABLE 13
Cornparison of acute and chronic toxicitv of test compounds to rainbow trout.
A
COIGOUND
LOG P
phenol
1.49
p-nitrophenol
1.91
p-chlorophenol
2.42
2,4-dichlorophenol
3.08
1,2,4-trichlorobenzene
4.26
pentachlorophenol
5.12
m TOXICITY
(UM)
REGRESSION ESTIMATE OF
CHRONIC TûXICITY
(UM)
RATIO OF
CHRONIC/ACUTE
[email protected] 1.1
Relationship between loq P and the loqarithas of acute and chfünic toxicity estimates iin un) üf the test cheiicalsa.
Acute:
Log LC50
Chrüniib: Log T
=
-
0.513 log
P + 2.733
6
0. 09
-5.58
31.1
0.29
O. 89
= - 0.507 log P + 1.713
nl nilaber of itiservatiin; Sb, standard deviation of the siope; tb, t-test of significance üf the slope; Fr, F-test of
signifieance of t h e rrgfessiün; S, standard errür o f the estimate; rZl coefficient o f determination.
Al 1 slopes and regfessions are statistically significant at the 0.05 prab~bility level.
Chïonie toxicity tbiest~oldsi T ) früa regression estinates.
significant, 25 percent reductions
in larval weight at hatch and at
10-days post-hatch respectively.
Thresholds based on fry weight
reduction
would
have
been
comparable to those based on larval
weight
decreases,
but without
the
associated
statistical
aignificance. Weight reduction of
larvae at hatch persisted to the
fry stage for al1 six test
chemicals. There was no recovery
of weight gain by the end of the
experiment as was observed in a
study of the effects of dioxin on
rainbow trout early life stages
(Helder 1980).
The criterion of a 25% change
is purely arbitrary and follows
current practice in the United
Sates (US EPA
1989).
This
criterion however, is not equivalent to environmental protection,
nor is it suitable for al1
parameters.
A good example is
percent moistulre. There were tests
in which percent moisture changed
significantly but deviated £rom
control values by only 1 to 3 %
(gag. phenol, Table 4 1 ,
Since
percent moisture reflects both
osmoregulation and the metabolism
of lipid, small changes are
important and may lead to other,
larger effects (e.g. changes in
percent mortality).
A more realistic approach to
thresholds would be to link them to
the natural variability of the
parameter. From Table 2, it is
evident that the range among
experiments for percent moisture of
larvae at hatch is 3.2%, while the
range for percent hatch is 11.5%
and percent mortality of eggs is
22.1%.
Clearly these variables
should not be treated identically.
From an ecological perspective, a 25% change in growth rate
may have effects on biomass
production
that vary widely,
depending on ecological conditions.
The significance can only be
assessed through estimates of
population performance, perhaps
through mcdelling.
For percent
mortality, a 25% increase for a
pericd of 30 days of fry growth
will lead to extinction of the
population if the mortality rate
persists.
Therefore, while a 25% change
£rom control represents a convenient and cornmon end-point, it is
dangerous to use it for decisions
on permissable chemical levels in
Development of water
water .
quality criteria £rom these data
will require a significant exercise
of judgement, and not a simple
mathematical
calculation.
Similarly,
interpolating
the
toxicity of non-tested compounds
£rom the QSAR for chronic toxicity,
creates significant risk of under
or over-estimation of potential
effects.
The logarithm of chronic and
acute
toxicity
was stronqly
correlated to the logarithm of the
octanol/water partition coefficient
(rZ = 0.75 and 0.89 respectively),
but the ratio of cases to independent variables was only 6 for this
study (a bare minimum requirement
is a ratio of 5 ) and large rZ
values may be artefacts.
The
regressions
and
slopes were
considered significant (p 5 0.051,
and the slopes were homogenous or
equal (p 5 0.05), indicating that
chronic
toxicity
for
these
chemicals could be
reasonably
predieted on the basis of acute
toxicity.
Although the slopes
appear identical (Figure 8), the
greatest deviation about the two
QSARs occurs at a log P of 3.08,
Since this partition
for W .
coefficient is mid-range, variations in measured toxicity do not
seriously affect the slopes of
QSARs
.
It is difficult to test the
assumption that cornmon modes of
action are the basis for QSARs,
i.e. that separate QSARs can be
developed
for each physical,
reversible mechanism of toxic
The
action (Arnold et al. 1 9 9 0 ) .
compounds tested in this study can
be classified in three different
groups: non-polar narcotics (TCB),
polar narcotics (PH, NP, 8,
DB,
P B ) and uncouplers of oxidative
phosphorylation (PB).
Narcosis by nonpolar compounds
is thought to represent a reversible wslowing"
of cytoplasmic
activity in the ce11 due to
chemical partitioning into the
biophase (Schultz 1989).
Schultz et al. (19861, workinq
with a heterogenous series of
phenols,
demonstrated
two
log P-dependent QSARs; one for
polar narcotic chemicals and the
other for "weak acidffuncouplers of
oxidative
phosphorylat ion.
Uncouplers usually contain two
nitro substituents or chlorine
substituents in the 4 or 5
position, as is the case for P B .
P B uncouples oxidative phosphorylation (Weinbach 1957) so that less
adenosine triphospate ( ATP) is
available for normal metabolic
processes.
To compensate, more
energy is
directed
to ATP
production which increases hsal
metabolism and reduces energy
available for growth.
This reduced growth efficiency
should be evident in reduced yolk
conversion efficiencies. P B and
TCB were the only compounds that
caused statistically significant
reductions in yolk conversion
efficiency.
Although decreases
were noted for the other compounds,
variability was too hiqh to
distinquish significant effects.
Therefore, the QSARs presented
here represent a mixed model, in
that several mechanisms of toxicity
Nevertheless,
may be acting.
strong relationships were observed.
Since log P mimics the partitioninq
of chemicals £rom water into lipid,
QSARs based on log P suqqest that,
for these compounds, the kinetics
of chemical accumulation are more
important determinants of toxicity
than specific mechanisms. Nevertheless, specific mechanisms m y
explain sorne of the variability
about the regressions.
Ionization of the phenolics in
water may also contribute to
QSAH
var iability
about
the
regression.
Saarikoski and
Viluksela
(1981) showed that
ionization of phenols and toxicity
varied with pH, presumably because
pH controls the relative proportion
of the more polar ion of each
phenol, which is taken up at a
lower rate the non-polar parent
compound
.
We have assumed that the
expressed
toxicity
of
these
chemicals was a function of a
steady-state exposure, where levels
of contaminants at the site of
toxic
action
are
virtually
constant.
While this may be the
case for the phenols, which rapidly
reach equilibrium in tissue ( < < 4
da-,
McCarty 1990),
Oliver and
Niimi (1983) demonstrated that the
rate of accumulation of chlorobenzenes by trout decreased with an
increasing degree of chlorination.
For TCB, equilibrium vas reached in
20 days, well within the 85-day
span of this expriment.
The average ratio of chronic
to acute (96 hr LC50) toxicities
was 0.13.
A second estimate of
this ratio may be calculated £rom
the ratios of the antilogs of the
intercepts of the two QSARs
(0.095).
However,
individual
ratios varied considerably and DB
(ratio = 0.02) differed most £rom
the overall mean.
The DB
threshold
concentration
mY
represent an extreme case since its
chronic toxicity is much greater
than might be expected on the basis
of its acute toxicity. The chronic
threshold concentration for DCP was
a low 0.32 p, hsed on growth, but
was consistent with the threshold
concentration
of 0.91pM for
chronic mortality. Three- to fivefold differences among laboratories
in
the
threshold
concentrations derived for the same
species and for the same test
chemical are typical (Woltering
1984). Nevertheless, the ratio can
only be considered in error if we
accept the assumption that al1
test chemicals have the same modes
of acute and chronic toxic action,
and hence the same ratios. We have
no evidence with which to test this
assumption.
The average ratio and the
intercept of the chronic toxicity
QSAR are also not fixed, i.e. they
will vary according to which
response is chosen as the basis for
the QSAR and what the criterion is
for a significant response.
The data published here were
first summarized as a preliminary
QSAR analysis by McCarty et al.
(1985). Their QSAR for the chronic
toxicity of chlorophenolics was not
parallel to their QSAR for acute
toxicity,
in contrast to our
analysis. The differences in the
QSARs are a function of the way the
data were analyzed:
1. The data were scrutinized for
errors much more carefully in
this study;
2, We estimated thresholds only
for those
parameters that
changed £rom control by more
than 25 %;
3. McCarty et
al.
(1985) used
Duncan's New Multiple Range Test
to identify treatment means that
were di£ferent £rom cont~rol.
This test is less reliable than
Dunnettfstest due to indeterminant experimental error (Day
and Quinn 1989);
4. The chronic threshold
tions were estimated
e t al. (1985) usinq
LOEC
approach;
reqression analyses;
concentraby McCarty
the NOECwe
uçed
5. McCarty e t
al.
(1985) used
several end-points to calculation each threshold; we consistently used growth of fry to
estimate thresholds;
6. The QSAR
in this analysis
included NP and TCB where as
McCarty et a l . (1985) included
only chlorophenolics.
was
In summary, a QSAR
described for the chronic toxicity
of orqanic chemicals to fish. This
QSAR was parallel to that for acute
toxicity.
The difference between
their intercepts was about one log
unit, indicating an average ratio
of chronic to acute toxicity of
about 0.10.
We gratefully thank Mike
Comba, Steve Munqer, Frank Varga,
and
Mike Brander,
Jane Kemp,
Wayne Jerome for their dedicated
technical support of this project.
Arnold, L.M., T.D. Lin and T.W.
Schultz.
1990.
QSAR for
and/or
chloromethyl
substituted anilines and the
polar narcosis mechanism of
toxicity.
183-191.
Chemosphere 21:
Bobra, A.M.,
D. Shiu,
and
D. Maekay. 1984. Structureactivity relationships for
toxicity
of hydrocarbons,
chlorinated hyùrocarbons and
oil
to
Daphnia
magna.
P. 3-16.
In: K.L.E. Kaiser
ied.1 QSARs in kivironmental
Toxicology.
D. Reidel
Publishinq Company, Dordrecht,
Ho1land ,
Day, R.W., and G.P. Quinn. 1989.
Cornparisons
of treatments
after analysis of variance in
Ecol. Mono. 59:
ecoloqy.
433-463.
Eaton, R.C.
and
R.O. Farley.
Growth and the
1974.
reduction of depensation of
zebrafish, Brachydanio rerio,
reared
in the laboratory.
Copeia 1974: 204-209.
Hansch, C., and A.J. Leo. 1979.
Substituent
Constants for
Analysis
in
Correlation
Chemistry and Biology. John
Wiley & Sons, New York, NY.
Helder, T. 1981. Effects of 2 , 3 ,
7,8 tetrachlorodibenzo-iioxin
{TCDD) on early life stages of
rainbow
trout
(Salmo
gairdneri ) .
Toxicol. 19:
101-112.
Hesselberg, R.J. and J.G. Seelye.
1982.
Identification of
organic compounds in Great
Lakes fishes by gas chromlography/mass
spectroscopy:
1979.
Great Lakes Fishery
Ann Arbor,
Laboratory,
Michigan.
Admin.
Rep.
No 82-1. 49 p.
and
B.R. Blunt.
Hodson, P.V.,
1981.
Temperature-induced
changes in pentachlorophenol
chronic toxicity to early life
stages of rainbow trout.
Aquat. Toxicol. 1: 113-127.
Hodson, P.V.
and
B.R. Blunt.
1986. The effect of time £rom
hatch on the yolk conversion
efficiency of rainbow trout
Salmo gairdneri.
J. Fish.
Biol. 29: 37-46.
D.G. Dixon,
and
Hodson, P.V.,
K.L.E. Kaiser.
1984. Measurement of median lethal dose
as a rapid indication of
contaminant toxicity to fish.
Env. Toxicol. Chem. 3:
243254.
Hodson, P.V.,
B.R. Blunt.
rainbow
gairdneri 1
exposure
selenium.
Aquat. Sci.
D.J. Spry
ad
1980. Effects on
trout
(Salmo
of
a chronic
to
waterborne
Can. J. Fish.
37: 233-240.
Jones, P.A. 1981. Chlorophenols
and their impurities in the
Canadian environment. Economic
and Technical Review, Report
EPS-3EC-81-2,
Environment
Canada.
Konemann, H.
1981. Quantitative
structure-activity relationships
in
fish toxicity
studies. Part 1: Relationships for
50 industrial
pollutants.
Toxicol. 19:
209-221.
Kwasniewska, K., and K.L.E. Kaiser.
1983. Toxicities of selected
phenols to fermentative and
oxidative yeast.
Bull. Env.
Cont. Toxicol. 31: 188-194.
Macek, K.J.,
and B.H. Sleiqht.
1977.
Utility of toxicity
tests with embryos and fry of
fish in evaluating hazards
associated with the chronic
toxicity of chemicals to
fishes.
P. 137-146.
In:
F.L. Mayer and J.L. Hamelink
[ed-1. Aquatic Toxicology and
Hazard Evaluation.
ASTM STP
634, PhiPadelphia, PA.
McCarty, L.S.
1990. A kineticsbased analyçis of quantitative
structure-activity relationships in aquatic toxicity and
bioconcentration
bioassays
with organic chemicals. PhD
thesis, Dept of Biology, U. of
Waterloo, 141 p.
McCarty, L.S.,
P.V. Hodson, G.R.
Craig
and
K.L.E.Kaiser.
1985. The use of quantitative
structure-activity relationships to predict the acute and
chronic toxicities of organic
Env
chemical to fish.
Toxicol. Chem. 4: 595-606.
.
McKim, J.M.
1977. Evaluation of
tests with early life stages
of
fish
for predicting
long-term toxicity. J. Fish.
Res. Board C'an. 34: 11481154.
McKim, J.M.
1985.
Early life
stage toxicity tests. P. 5895. In: G.M. Rand and S.R.
Petrocelli led.1.
Fundamentals of Aquat ic Toxicology.
Hemisphere Publishing Co.
Mount, D.I.,
and
W.A.Brungs.
1967.
A simplified dosing
appéiratus
for
toxicology
studies.
Water Res. 1:
21-29.
.
NAS/NAE
1973.
Water Quality
Criteria 1972.
National
Academy of Sciences, National
Academy
of
Engineering.
Washington, D.C.
EPA. R3.
73.033. 594 p.
An
Norbert-King, T.J.
1988.
interpolation estimate for
chronic toxicity:
The ICp
approach.
Technical Report
05-88,
National
Effluent
Toxicity Assessment Center,
Environmental
Research
Laboratory, US Environnemtal
Protection
Minnesota.
Agency,
Duluth,
Oliver, B., and A. Niimi.
1983.
Bioconcentration of chlorobenzenes £rom water by rainbow
trout :
correlation with
partition coefficients and
environmental residues. Env.
Sci. Tech. 17: 287-291.
Saarikoski, J. and M. Viluksela.
1981. Influence of pH on the
toxicity
of
substituted
Arch
phenols to
fish.
Environ.
Contam. Toxicol.
10: 747-753.
.
Schultz, T.W.
1989.
Nonpolar
narcosis :
a review of the
mechanism
of
action for
hseline toxicity. P . 10. In
U.M. Cowgill and L.R. Williams
(ed.1. Aquatic Toxicology and
Environmental Fate: Twelfth
Symposium
ASTM STP-1027.
Philadelphia, PA.
.
Schultz, T.W., G.W. Holcombe, and
G.L. Phipps. 1986. Relationships
of
quantitative
structure-activity
to
comparative
toxicity
of
selected phenols in P i m e p h a l e s
promelas
and
Tetrahymena
p y r i f ormis
test
systems
Ecotox. Env. Saf. 12: 146153.
.
Schultz, T.W., and B.A. Moulton.
1984.
Structure-activity
correlations
of
selected
azaarenes, aromatic amines,
and nitroaromatics.
P. 337-
352.
In K.L.E. Kaiser led.1
QSARç
in
Environmental
Toxicology.
D. Reidel
Publishing Company, Dordrecht,
Holland.
Sokal, R.R., and F.J. Rohlf. 1981.
Biometrv.
2nd ed.
W.H.
Freeman and Co. San Francisco,
CA.
Sprague, J.B. 1973. The AB& of
pollutant bioassay using fish.
P. 6-30.
In:
Biological
Methods for the Assessment of
Water Quality. AÇTM STP 528.
Philadelphia, PA.
Steel, R.G.D.
and J.H. Torrie.
1960. Principles and Procedures of Statistics. McGrawHill Book Co. Inc., Toronto.
481 p.
Suntio, R.L., W.Y. Shiu, and D.
Mackay. 1988.
A review of
the nature and properties of
chemicals present in pulp mil1
effluents.
Chemosphere. 7:
1249-1290.
Suter, G.W.
1990.
Seven-day
tests and chronic tests.
Letter to the Editor. Env.
Toxicol. Chem. 9: 1435-1436.
U.S. Environmental
Protection
Aqency. 1980. Ambient water
quality criteria for chlorinated benzenes. EPA-440/5-80028. Washington, DC.
U. S. Environmental
Protection
Agency
1989.
Short-term
methods for estimating the
chronic toxicity of effluents
and
receiving waters
to
freshwater organisms 2nd ed
EPA-600/4-89-011. Washington,
DC.
.
.
.
Van Leuwen, C.J., D.M.M. Adema, and
J. Hermens. 1990. Quantitative
structure-activity
relationships for fish early
life stage toxicity. Aquat.
Toxicol. 16: 321-334.
Veith, G., D. Call, and L. Brooke.
1983.
Structure-toxicity
relationships for the fathead
minnow, (Pimephales promelas ) :
Narcotic industrial chemicals
Can. J. Fish. Aquat. Sci. 40:
743-748.
.
Verschueren, K. 1983. Handbook of
Environmental Data on Organic
Chemicals
Van Nostrand
Reinhold Co. Inc., New York,
NY.
.
Weinbach, E.C.
1957. Biochemical
basis
for
toxicity
of
pentachlorophenol.
Proc
Natl. Acad. Sci. 43: 393-397.
.
Woltering, D.M. 1984. The growth
response in f ish chronic and
early life stage toxicity
tests :
A critical review.
Aquat. Toxicol. 5: 1-21.
Yalkowsky, S.H., R . J . O r r , and S.C.
Valvani.
1979.
Solubility
3.
The
and p r t i t i o n i n g .
s o l u b i l i t y of halobenzenes i n
Indust. Ing. Chem.
water
Fundament. 18: 351-353.
.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement