HAL Id: hal-00902446
https://hal.archives-ouvertes.fr/hal-00902446
Submitted on 1 Jan 1996
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of
sci-entific research documents, whether they are
pub-lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Water quality parameters associated with Aeromonas
spp-affected hatcheries
C Ortega, Jl Muzquiz, Arnaud Fernandez, I Ruiz, I de Blas, Mc Simon, Jl
Alonso
To cite this version:
Original
article
Water
quality
parameters
associated
with Aeromonas
spp-affected
hatcheries
C
Ortega
JL
Muzquiz,
A Fernandez
I Ruiz,
I
De Blas, MC Simon JL Alonso
Enfermedades Infecciosas y
Ep!demíolo!ía,
DepPatología
Animal, Facultad de Veterinaria,Miguel
Servet 177,
50013 Zaragoza, Spaí!
(Received
6April
1995;accepted
25April 1996)
Summary ―
Thisstudy investigates
the role of waterquality
parameters aspossible
risk orprotection
factors in Aeromonas
spp-affected
rainbow trout hatcheries in northeasternSpain.
The results revealedan association between oxygen concentration, ammonia concentration and total dissolved solids
(TDS)
with theprevalence
of Aeromonas spp cases on the affected fish farms. The oxygen andammo-nia concentrations acted as risk factors when their values were lower than 7.06
mg/L
orhigher
than 0.05mg/L, respectively, generating
odds ratio(OR)
values of 2.0 and 2.3. TDS levels, however, actedas a
protective
factor(OR
=0.39),
from anepidemiological point
of view but notstatistically,
when the values were outside the normal range, 199-261mg/L.
On a multivariate level, fish farmsexposed
to all the associated factors at the same time, had a risk ofbeing
affectedby
Aeromonas spp thatwas 1.74 times
higher
than farms with normalquality
parameters; the risk increased when TDS had anormal value and it decreased when oxygen and/or ammonia concentrations had normal values.
epidemiology
/ odds ratio /Oncorhynchus
mykiss
/ waterquality
/ AeromonasRésumé ―
Étude
du rôle desparamètres
dequalité
de l’eau dans lespiscicultures
de truite(Oncorhynchus mykiss)
affectées par Aeromonas spp dans le nord-est del’Espagne.
Nousavons effectué une étude sur le rôle des
paramètres
dequalité
de l’eau, comme facteurspotentiels
derisque
ou deprotection,
enpiscicultures
affectées par Aeromonas spp. Les résultats ont démontré l’exis-tence d’une relation entre la concentrationd’oxygène, d’ammoniaque
et les solides totaux dissous(STD)
dans l’eau et laprésence
dAeromonas spp. Les concentrationsd’oxygène
etd’ammoniaque
ontagi,
comme des facteurs derisque quand
les valeurs étaient inférieures à 7,06mg/L
etsupérieures
a0,05
mglL,
respectivement.
Les valeurscorrespondantes
du Odds Ratio étaient de 1,97 et 2,6. En revanche les STDagissaient
comme un facteur protecteurquand
les valeurs se trouvaient hors de l’inter-valle considéré comme normal(dans
le milieu d’étude : 199-261mglL
avec un Odds Ratio de 0,39).Dans
l’analyse
multivariée nous avons observé que les pisciculturesexposées
aux trois facteurs en*
même temps ont un
risque
1,74 foissupérieur
d’être infectéespar Aeromonas
spp que celles nonexposées
à ces facteurs simultanément. Lerisque augmentait quand
les valeurs de TDS étaientnor-males et diminuait
quand
les valeurs de la concentrationd’oxygène
ouletd’ammoniaque
étaientnor-males.
épidémiologielOdds Ratio/Oncorhynchus mykisss
/ qualité
de 1 &dquo;eau / AeromonasINTRODUCTION
Disease outbreaks due to Aeromonas spp,
especially
in the case ofFurunculosis,
where Aeromonas salmonicida is theaetiological
agent,
cause the most serious economic losses inSpanish
salmonid hatcheries(Ortega
etal,
1993a,b;
Sanzet al,
1993).
Together
with Asalmonicida,
there are other Aeromonas sppspecies
such as Ahydrophila,
Asobria,
Acaviae,
which areassociated with disease outbreaks in
aqua-culture,
in bothextensively
andintensively
cultured fish(Paterson,
1983; Barja
andEst6vez, 1988).
All theseaetiological agents
are
present
in a latent state on alarge
num-ber of fish
farms,
oftenproducing
asymp-tomatic carriers(Jensen
andLarsen,
1985;
Austin andAustin, 1987).
Aeromonas spp have been detected in salmonid andnon-salmonid fish
species,
in both saltwater andfreshwater,
and causehigh
mortality
rates(Cornik
etal, 1984;
Austin andAustin,
1987;
Nougayred
etal, 1990; Wiuklund,
1990).
Disease outbreaks affect bothgrowing
fish andadults,
and the clinicalpresentation
ofthe diseases is characterized
by
suchsigns
as the presence of lesions
resembling
boils in the skeletalmusculature,
exophthalmus,
lethargy,
blood-shotfins,
and/or darker skin colour(De
Kinkelin etal, 1985;
Austin andAustin, 1987;
Barja
andEst6vez, 1988).
Diseasesproduced
by
Aeromonas spphave been detected in association with envi-ronmental factors such as water
tempera-ture, season of the year, or with
biological
factors such as the
immunological
statusand stress level of the fish
(Michel
andDubois-Darnaudpeys,
1980;
Ohtsuka etal,
1984; Thorburn,
1987; Bly
etal,
1993;
Jarp
et
al, 1993;
Whipple
andRohovec,
1994;
Ortega
etal,
1995).
Control of these factors couldhelp
to reduce the effects of diseaseoutbreaks on fish farms and could
provide
better information about the role ofasymp-tomatic carriers in these disease outbreaks
(Jensen
andLarsen,
1985;
Ortega
etal,
1995).
The
objective
of thisstudy
was toevalu-ate the role of water
quality
parameters
asrisk factors on fish farms affected
by
Aeromonas spp at a univariate and multi-variate level. This could thenhelp
bettercontrol these
damaging aetiological
agents,
which are the most common inSpanish
aquaculture
(Sanz
etal,
1993).
MATERIALS AND METHODS
The
study
was carried out in 12 hatcheries in northeasternSpain,
over a 3 yearperiod
(1992-1994).
All the fish farmsproduced
rain-bow troutfingerlings throughout
the year. The studied farms had cement andfiberglass
pools
depending
on the age of the fish. In all the farms,the water was
supplied by
rivers. At every farm watersamples (see
Sample
selection(fish))
and fish(see
Waterquality parameters)
from three differentpools
were taken every 3 months(one
sampling
occasion perseason).
Sample
selection(fish)
The number of sampled fish to be taken from every fish farm was determined
using
thecom-puter programe
Episcope (Frankena
et al,1990)
and the formulas for
sampling
selection inGraat,
1994).
We used an expected prevalence of 30%(the
value observedby
Sanz et al(1993)
for Aeromonas spp infection inprevious
studies inSpain),
an accepted error of 10% and a confi- i-dence level of 90%. This determined asample
size of 57 animals per
hatchery. Taking
intocon-sideration the
possibility
of lossesresulting
fromdiagnostic
and dataanalysis,
thesample
sizewas increased
by
5% to 60 animals for every fish farm. Thissample
included fish of all different ages and both sexesliving
on the farms,using
a likelihood selection method
(Martin
et al, 1987; Thrusfield, 1990;Ortega
and Graat,1994).
Aeromonas
spp-affected
hatcheriesA
hatchery
was considered to be affectedby
Aeromonas spp if the presence of symptomsassociated with Aeromonas spp was detected
(lesions
resembling
boils in the skeletic muscu-lature,exophthalmus, lethargy,
blood-shot fins,multiple haemorrhages
in musculature or darker skin colour(De
Kinkelin et al, 1985; Austin andAustin, 1987;
Barja
and Est6vez,1988))
and/or if Aeromonas spp were isolated from the organs of the fish in culture media.Aeromonas spp isolation
The
sampled
fish were sacrificedusing
ahigh
dose of
2-phenoxyethanol,
andsamples
weretaken from the
kidney,
liver,spleen
and lesions in skeletal musculature.Samples
were cultured intryptic
soy agar(TSA)
and brain heart infusion agar(BHIA),
at 22 °C for 48-72 h. Isolated bac-teria were identifiedby microscopy,
biochemical andstaining procedures (Austin
and Austin, 1987;Barja
and Est6vez, 1988).Water
quality
parameters
Water was
sampled
in three different locations(pools)
of each fish farm. Watersamples
weretaken from the centre of the
pools (10-20
cm indepth).
Thefollowing
measurements were madeusing
portable electronicequipment, colorimetry
reagents and a
portable spectrophotometer:
watertemperature
(°C), pH,
oxygen(mg/L),
ammonia(mg/L),
nitrate(mg/L),
nitrite(mg/L),
hardness(mg/L), alkalinity (mg/L),
carbon dioxide(C0
2
)
(mg/L),
chemical oxygen demand(COD) (mg/L),
conductivity
(
N
S/cm
2
),
total disolved solids(TDS)
(mg/L)
and oxygen reduction potential(REDOX)
(mV) (Stirling, 1985).
These parameters were the continuousindependent
variables of thestudy.
Epidemiological
and statisticalanalysis
The statistical
analysis
was carried out as across-sectional
study (Martin
et al, 1987; Thrusfield,1990),
where the association between Aeromonasspp-affected
fish farms and waterquality
param-eters wasanalyzed
at three different levels.The first level of determination was of a
pos-sible statistical association between water
quality
parameters and Aeromonas
spp-affected
fish farms. At this level,only
factors with a P value lower orequal
to 0.25 were selected, which meant the factors morelikely
to be associated with Aeromonasspp-affected
fish farms.The second level determination used the odds ratio
(OR),
of thequantitative
value of theasso-ciation between Aeromonas
spp-affected
fish farms and waterquality
parameters at an uni-variate level(role
of every associated factor eval-uatedindependently).
This second levelrequired
the transformation of the continuous parameters into dichotomic parameters. This was madeby
defining
cut-offpoints
for every parameter(Frankena
and Thrusfield,1994) giving
two pos-sible results for every parameter, normal valuesor abnormal values. Values were considered
nor-mal if
they
fell within the confidence interval(CI)
defined
by
the parameters in thestudy
areaaccording
to thefollowing
formula:where X= = mean; SD = standard deviation; n =
population
size; Z= value of Student’s ttest at adefined confidence level
(Z=
1.96 at a 95%con-fidence level for two-tailed
distributions).
This defined two cut-offpoints
for every parameter: the lower and the upper CI limits.The third level determination used the OR of the quantitative value of the association betweenAeromonas
spp-affected
fish farms and waterquality
parameters at a multivariate levelA cross-sectional
study
wasadopted
rather than a cohort or case-controlstudy,
because every farm wasstudied
12 times. The animals and the factor values, however, were different every time, so everysampling
visit at agiven
farm had to be treated as a different case. The data base and statistical
analysis
of the factorswere made
using
the computer program EPI-INFO 5.0(WHO)
and OR was determinedusing
the
Episcope
program(Frankena
et al,1990).
The statistical association between disease and factors was determined
using
the Student’s ttestor
chi-square (depending
on the type ofvariable)
and P value. Thequantitative
value of theasso-ciation was determined
by
transformation of the continuous variables into dichotomic variables anddoing
the calculation of the OR and itscon-fidence interval
(CI) (Frankena
and Thrusfield,1994).
P means the likelihood of a real association between the disease and the factor. An associa-tion with P values < 0.25 was
accepted
for thepreliminary
factorscreening.
OR means the
quantitative
value of the asso-ciation, and itsinterpretation
was ’animalsexposed’
to the variablehaving
an ’OR value’indicating
ahigher
risk for disease thannon-exposed animals. When OR = 1 association did
not exist; OR < 1 the factor acted as a ’protective’;
OR > 1 the factor acted as a ’Risk’.
CI means
accepted
limits for the OR. When the value 1 was included in the interval, amea-surable association at a
quantitative
level did notexist; an interval lower than 1 indicated a
’pro-tective factor’; an interval
higher
than 1 indicateda ’risk factor’.
RESULTS
The
study
determined that theprevalence
of Aeromonasspp-affected
fish farmsduring
the overallstudy
period
was 19.5% innorth-eastern
Spain.
It was observed that theprevalence
of Aeromonasspp-affected
fishfarms decreased
during
the 3 years of thestudy:
24.0% in1992,
18.7% in 1993 and16.7% in 1994. The results of the associa-tion between Aeromonas
spp-affected
fish farms and waterquality parameters
at the three studied levels are discussed below.Determination of associated
parameters
In the statistical
analysis
of the association of theparameters
with the Aeromonas sppstatus of the fish farms at a 75% of confi-dence level
(P
< 0.25),
apossible
statisti-cal association wasonly
detected forpH,
oxygen,
TDS,
conductivity, ammonia,
nitrate,
hardness and
alkalinity
(see
tableI).
Theseresults of the first level of the
study
were apreliminary
screening
of the factors thatwere most
likely
to be associated with Aeromonasspp-affected
fish farms. The real role of these associatedparameters
was
quantified using
OR in the second level of the dataanalysis.
Quantitative value of the association
between water
quality parameters
and Aeromonas-affected fish farms at univariate level
At the second level of the
study,
cut-offpoints
for normal values of the waterquality
parameters
were determined(table II).
Val-ues
falling
outside these limits werecon-sidered abnormal. In the cases of oxygen,
ammonia and
nitrate,
only
one CI limit wasconsidered as the cut-off
point,
because it is known that low oxygen andhigh
ammonia and nitrate concentrations are harmful tosalmonid
species.
The
quantitative
values of the associa-tion were determinedusing
the OR and its CI as shown in table II. The oxygen and ammonia concentrations were risk factors associated with Aeromonasspp-affected
fish farms because the Cls weregreater
than one. While TDS was not
statistically
associated as a
protection
factor,
becauseone was included in the
CI,
it could becon-sidered as a
protection
factor from anepi-demiological point
of view(the
upperlimit,
1.08,
was so close to one and the Cl wasidentified as
being
associated with Aeromonasspp-affected
fishfarms,
the Clswere too wide to define the
quantitative
value of the association. All of them includedone well within their ranges.
Quantitative value of the association between water
quality
parameters
and Aeromonas-affected fish farms
at multivariate level
The calculated OR
represented
the role of each associated factor at a univariate level. Thepossibility
that the factors interacted at a multivariate level was examinedusing
stratification,
but in all cases the result wasnegative.
The role of these factors at a multi-variate level issimply
theproduct
of the indi-vidual OR values when interaction factors do not exist.The multivariate
analysis
results revealed that the risk of fish farms where the oxygenand ammonia concentrations have abnor-mal values
(risk factors) being
anAeromonas
spp-affected
fish farm was 4.45higher
than in a farm with normalvalues,
while if the three associated factors
together
were included in the
analysis (including
the TDS abnormal values as aprotective
fac-tor),
the total risk was 1.74 timeshigher
thanin fish farms where these factors were not
present
together.
DISCUSSION
The observed
prevalence
of Aeromonasspp-affected
fish farms inSpain (19.5%)
was lower than that observed 3 years ago
(31 %)
(Sanz
etal,
1993),
and it was also lower than the detectedprevalence
of 24% in the same studied area in 1992(Ortega
et
al,
1993a).
We also observed that preva-lence of Aeromonasspp-affected
fish farms decreased over thestudy period (24%
inThe observed results
implied
asignifi-cant reduction of
prevalence
which wasprobably
due to the fact that at thebegin-ning
of thisstudy
we decided to set up newcontrol
systems
inhatchery management
whenhigh
prevalence
of the disease wasdetected. The disease reduction would not
be
possible only
as a ’time effect’ without these newmanagement systems (these
newmanagement systems
wereapplied
before to the
begining
of thestudy
and the effect of thesemanagement
systems
wasobserved
during
ourstudy).
With
regards
to thestudy
of waterqual-ity
parameters,
oxygen concentrations below 7.06mg/L
and ammonia concentrations above 0.05mg/L
are risk factors associated with Aeromonasspp-affected
fish farms in northeasternSpain
at a 95% confidence level. Fish farms with waterquality
which exceeded these limits were about twice aslikely
to suffer from Aeromonas spp. Thestudies of other authors also determined a
role of oxygen and ammonia concentration
as factors associated with disease
preva-lence in Atlantic salmon
(Salmo
salarL),
but the role of these factors was modifiedover
time,
depending
on other factorsinter-acting
with oxygen and ammonia such asenvironmental factors
(season,
watertem-perature
and waterpH)
ormanagement
fac-tors
(intensification
in theproduction
andfeed) (Bergheim
etal,
1991;
Fivelstad,
1991
These
interactions were not detected in ourstudy
in the case of environmentalfactors
using
stratificationanalysis
and in the case ofmanagement
we did notstudy
these factors because there were no
sig-nificant differences in
management
between the studied farms. In contrast to ourresults,
other authors did not find any association
between oxygen and ammonia
concentra-tion and disease occurrence, but
they
foundan association of some fish
diseases,
includ-ing
diseases causedby
Aeromonas spp with watertemperature
andpH
in salmonid(Whipple
andRohovec, 1994)
andnon-salmonid fish
species
(Bly
etal,
1993).
Another
interesting
result was that TDSconcentrations outside the CI considered normal in our
study
area(199-261 mg/L)
appeared
to be aprotective
factor from anepidemiological
point
of view but a statistical association was ’not’ detected because 1 was included in the Cl of the OR. Fish farmswith such values were
only
39% aslikely
tosuffer from Aeromonas spp infections as
fish farms with normal values.
TDS and
conductivity
which are oftenclosely
related. We also did not detect any interac-tion between these two factors in the
strat-ified statistical
analysis.
We think thisinter-action exists because the Cl of the OR for the
conductivity
had a broad range(a
broadrange in the CI means interaction
(Martin
et
al, 1993;
Thrusfield,
1990)
orconfound-ing).
Perhaps
it was not detected becausean insufficient number of water
samples
were testedduring
thestudy
(economical
reasons made it
impossible
to testmore).
It ispossible
that normal values of TDSact as risk factors because
they
couldplay
some role in Aeromonas spp
activity
ormul-tiplication.
These normal values could beoptimal
for the normalactivity
of Aeromonas spp, while the abnormal values observedin our
study
area, both above and belowCI,
could deter bacterialactivity
but not fish life. It has then to be considered that thecon-ductivity
couldplay
some similar role to that ofTDS,
because a close relation has to exist between both waterquality parameters.
Thiswas
suspected
but not determined in thestudy.
We did not detect any association
between
temperature
and Aeromonasspp-affected fish farms as has been observed
by
other authors(De
Kinkelin etal, 1985;
Austin and
Austin, 1987;
Barja
andEst6vez,
1988).
Thismight
be because we did notconsider as affected fish farms
only
thosefarms with Furunculosis
symptoms
andmor-tality
but we also included the farms onwhich we isolated Aeromonas spp. That means we detected both fish farms with the disease
(Furunculosis,
which isclosely
related totemperature)
and fish farms with latent Aeromonas spp infection(which
isnot
closely
related totemperature).
The multivariate effect of
combining
thesignificant
risk factorstogether
indicated that farms withabnormally
low disolvedoxy-gen levels or
high
ammonialevels,
but with normalTDS,
were about 4.45 times aslikely
to suffer from Aeromonas spp infections.The presence of abnormal TDS values
decreased the risk to about
1.74,
which is lower than when oxygen or ammoniacon-centrations act as risk factors alone.
It seems
possible
that the risk level could increase withmanagement factors,
as wasobserved in recent studies of risk factors
associated with A salmonicida infection where it was determined that environmental
parameters acting together played
a lessimportant
role thanmanagement
inFurun-culosis outbreaks
(Jarp
etal,
1993)
ACKNOWLEDGMENTS
We would like to thank JL Gimenez and M Lamuela. The govemment of
Arag6n
and Navarrasupported
theproject economically.
REFERENCES
Austin B, Austin DA (1987) Bacterial Fish Pathogens, Disease in Farmed and Wild Fish. Ellis Horwood Limited, Chichester, 111-177, 308-329
Barja JL, Estdvez A (1988) Patologia en acuicultura. Enfermedades bacterianas de los peces. Espinosa
y Labarta, Madrid, 485-490
Bergheim A, Seymour EA., Sanni S, Tyvold T (1991)
Measurements of oxygen consumption and
ammo-nia excretion of Atlantic salmon (Salmo salar L) in commercial-scale, syngle-pass freshwater and
sea-water landbased culture systems. Aquacult Eng 10, 251-267.
Bly JE, Lawson LA, Szalay AJ, Clem LW (1993) Envi-ronmental factors affecting outbreaks of winter
sapro-legniosis in channel catfish, Ictalurus punctatus
(Rafinesque). J Fish Dis 16, 541-549
Cornik JW, Morrison CM, Zwicker B, Shun G (1984)
Atypical Aeromonas salmonicida infection in atlantic cod, Gadus morhua L. J Fish Dis 7,495-499 De Kinkelin P, Michel C, Ghittino P (1985) Precis de
pathologie des poissons. INRA-OIE, Paris, 94-116 6
Fivelstad S (1991) Studies of limiting factors governing the water flow requirement for Atlantic salmon (Salmo
salar L) in landbased seawater systems. Aquacult
Eng 10, 237-249
Epidemiology Seminar. CIHEAM- Zaragoza, 17-28 October, ch 5, 1-35
Frankena K, Noordhuizen JP, Willeberg P, Van
Voor-rthuysen PF (1990) Episcope: computer programs in
veterinary epidemiology. Vet Rec 9, 573-557
Jarp J, Tangen K, Willunsen FV, Djupvik HO, Tveit AM
(1993) Risk factors for infection with Aeromonas salmonicida subsp salmonicida in Norwegian fresh-water hatcheries. Dis Aquat Org 17, 81-86 Jensen NJ, Larsen JL (1985) Seasonal occurrence of
Aeromonas salmonicida carriers. In: Fish and Shell-fish Pathology (AE Ellis, ed), Academic Press, Lon-don, 87-89
Martin SW, Meek AH, Willeberg P (1987) Veterinary Epidemiology, Principles and Methods. Iowa State
University Press, Ames, 10, 149-201
Michel C, Dubois-Darnaudpeys A (1980) Persistence of the virulence of Aeromonas salmonicida strains
kept in river sediments. Ann Rech Vet 11, 375-380
Nougayred P, Sochon E, Vuilaume A (1990) Isolation of Aeromonas salmonicida in farmed turbot (Psetta
maxima ) in France. Bull Eur Assoc Fish Pathol 10, 139-140
Ohtsuka H, Nakai T, Muroga K, Jo Y (1984) Atypical Aeromonas salmonicida isolated from diseased eels. Fish Pathol 19, 101-107
Ortega C, Graat L (1994) Principles and methods of
sampling in animal disease surveys. Models and Quantitative Methods in Veterinary Epidemiology Seminar. CIHEAM, Zaragoza, 17-28 October, ch 3, 1-24
Ortega C, Muzquiz JL, Sanz F, Ramos P, Planas E, Garcia J, Docando J (1993a) Situaci6n actual de la
patologia en acuicultura continental de salm6nidos
en la peninsula lb6rica: II. Estudio virol6gico. World
Aquaculture Congress Proceedings. Torremolinos,
Spain, 21-23
Ortega C, Muzquiz JL, Planas E, Docando J, Alonso JL, Simon MC (1993b) Furunculosis seroprevalence in continental aquaculture in Northeast Spain.
Pre-liminary study. Vlth International Conference on
Dis-eases of Fish and Shellfish Proceedings, Brest, France, EAFP, Brest, 34-35
Ortega C, Muzquiz JL, Docando J, Planas E, Alonso JL, Simon MC (1995) Ecopathology in aquaculture: risk factors in infectious disease outbreak. Vet Res 26, 57-62
Paterson WD (1983) Furunculosis and other associated diseases caused by Aeromonas salmonicida. In:
Antigens of Fish Pathogens (DP Anderson, M Dor-son, PH Dubourget, eds), Collection Fondation Mar-cel Merieux, Lyon, 119-137
Sanz F, Ramos P, Ortega C, Muzquiz JL, Planas E, Girones 0 (1993) Situaci6n actual de la patologia
en acuicultura continental de salm6nidos en la
penin-sula lb6rica; i. Estudio bacteriol6gico y parasitol6gico.
World Aquaculture Congress Proceedings. EAS, Torremolinos, Spain, 24-25
Stirling HP (1985) Chemical and Biological Methods of Water Analysis for Aquaculturists (MC Beveridge, LG Ross, MJ Phillips, eds), Piscis Press,
Univer-sity of Stirling, Stirling, UK
Thorburn MA (1987) Factors influencing seasonal vib-riosis mortality rates in Swedish pen-reared rainbow
trout. Aquaculture 67, 79-85
Thrusfield M (1990) Estudios observacionales. In:
Epi-demiologia Veterinaria. Acribia SA, Zaragoza, 207-217 7
Whipple MJ, Rohovec JS(1994). The effect of heat and low pH on selected viral and bacterial fish pathogens.
Aquaculture 123, 179-189