Reviews in Aquaculture (2020) 12, 524–541 doi: 10.1111/raq.

12331

A systematic literature review of the major factors causing

yield gap by affecting growth, feed conversion ratio and
survival in Nile tilapia (Oreochromis niloticus)
Samuel Bekele Mengistu1,2,3 , Han A. Mulder1 , John A. H. Benzie2,4 and Hans Komen1
1 Wageningen University & Research, Animal Breeding and Genomics, PB Wageningen, The Netherlands
2 WorldFish, Jalan Batu Maung, Batu Maung, Bayan Lepas, Penang, Malaysia
3 School of Animal and Range Sciences, College of Agriculture, Hawassa University, Awassa, Ethiopia
4 School of Biological Earth and Environmental Sciences, University College Cork, Cork, Ireland

Correspondence Abstract
Samuel Bekele Mengistu, Animal Breeding and
Genomics, Wageningen University & Research, Productivity among small- and medium-scale tilapia farms varies considerably.
Droevendaalsesteeg 1, 6708 PB Wageningen, The difference between the best performers and lower ones (yield gap), is affected
The Netherlands. Emails: by differences in growth rate and feed conversion ratio (FCR). FCR at the farm
[email protected]/ level is strongly influenced by survival of fish. In this study a systematic literature
[email protected]
review of two databases (ASFA and CAB-Abstracts) identified 1973 potentially
relevant articles. Data from 32 articles that met the inclusion criteria were anal-
Received 10 October 2018; accepted 5 Febru-
ary 2019. ysed using linear mixed models for the most important factors with significant
contributions to growth [investigated through analysis of the thermal growth
coefficient (TGC)], survival and FCR of Nile tilapia. Increasing crude protein
(CP), dissolved oxygen (DO) and pH significantly decreased FCR and increased
TGC. Increasing stocking weight (SW) significantly improved both FCR and sur-
vival. Temperature had the largest effect on FCR followed by DO, pH and CP.
DO had the largest effect on TGC followed by CP and pH. This study confirms
that the optimal rearing temperature for Nile tilapia is between 27 and 32°C.
Improving management to optimize DO (> 5 mg/L), stocking density (3–5 fish/
m2), SW (> 10 g) and CP (25
30%) will improve performance and survival in
small- and medium-scale tilapia farming. However, it is hard to influence temper-
ature in ponds and cages while DO is largely influenced by aeration. Since many
small- and medium-sized farms do not have aeration, these major tilapia farming
systems could benefit from genetically improved strains selected for resilience to
highly fluctuating diurnal temperature and DO levels.
Key words: feed conversion efficiency, growth, survival, tilapia, yield gap.

systems ranging from artisanal to intensive systems (Nor-

Introduction
man-L
opez & Bjørndal 2009). Tilapia is an important fish
The supply of fish for human consumption has been species for home markets in Asia, South America and
increasing at a rate of 3.2% per year since the 1970s until Africa; the United States of America is the major export
2013. Aquaculture made a substantial contribution to this market for tilapia (FAO 2016). Therefore, many selective
increase, with inland finfish farming contributing 65% of breeding programs have been established for Nile tilapia
the increase in fish production from 2004 to 2014 (FAO (Neira 2010) including an important non-commercial
2016). Among the finfish, Nile tilapia (Oreochromis niloti- breeding program by WorldFish that developed the
cus) ranked second in terms of production volume next genetically improved farmed tilapia (GIFT). The GIFT
to carps (grass carp, silver carp and common carp) with strain has been disseminated to many countries (Komen
a total production volume of 3.7 million tons worth & Trinh Quoc 2014). According to Neira (2010), 10 out
about 6 billion USD (FAO 2016). Nile tilapia is farmed of 17 Nile tilapia breeding programs had used the GIFT
in more than 80 countries and in different production strain as their base population.

524 © 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd
This is an open access article under the terms of the Creative Commons Attribution License, which permits use,
distribution and reproduction in any medium, provided the original work is properly cited.
 Determinants of yield gap in Nile tilapia farms

Nile tilapia production systems can be classified in terms et al. (2013) reported a mortality rate of 71–72% for the
of input utilization as extensive, semi-intensive and inten- cage culture environment, 48% for the pond nucleus envi-
sive farming systems. The earthen-pond production sys- ronment and 32% in the polyculture production environ-
tems are the dominant ones practiced by small- and ment in Vietnam. The economic effect of mortality
medium-sized tilapia farms. Such farms typically produce depends on the stage during which fish mortality happens.
fish of 200–500 g weight targeting local markets. Larger fish Mortalities occurring during the later stages of the grow-
with harvest weights above 800 g are produced by large out phase have the largest economic impact due to the
farms that mostly use larger ponds with aeration, or cages accumulated cost of production. The amount of feed deliv-
in lakes and reservoirs (Omasaki et al. 2016b, Hoong Yip ered at any one time is usually based on the estimated
Yee, pers. comm., 2016). Currently there is a big difference standing stock of fish and the FCR is measured based on
in productivity among many small- and medium-sized tila- the amount of feed fed and the biomass harvested. Overes-
pia farms. The difference in productivity between the best timating the standing stock will increase the feed waste,
performing farms and low performing farms is defined as a which has a negative effect on profit and environment,
yield gap, the difference between achieved production and while underestimating leads to underfeeding of the fish and
that which is possible with optimal management. Many fac- reduced production.
tors can contribute to differences in productivity but all The wide range of FCR and mortality values reported
have their action ultimately in their effects on growth, sur- indicate a large difference between the best and worst per-
vival and feed efficiency, and this can be summarized as a forming farms and suggest significant room for improve-
difference in feed conversion ratio (FCR). There are large ment with respect to more efficient husbandry. The
differences in FCR and survival among many small- and investment in genetic improvement programs designed to
medium-sized tilapia farms. improve performance in farming systems is undermined by
FCR at the level of production units is defined as the these inefficiencies. Investigating the factors that contribute
ratio of the total feed given divided by total biomass har- to the reduced productivity of tilapia fish farms is critical
vested. FCR is determined by individual feed efficiency to providing the information needed to tackle the yield gap
and survival, because fish that die during the grow-out problem. First, by determining whether husbandry
period eat feed until death, but do not contribute to the approaches can be optimized, and second, for those aspects
total biomass harvested. Reported FCR values for tilapia of the environment that cannot be managed, identifying
vary widely, ranging from 1.5 to 2.5 in pond environ- whether farmed strains can be genetically improved to be
ments and from 1.0 to 1.71 in cage environments (Rana more efficient in those environments.
& Hassan 2013). Thoa et al. (2016) reported FCR values Work over the last two decades has established some of
of 1.08 and 1.89 in freshwater and saline water pond the main parameters for optimizing the environment for
environments, respectively. FCR is considered acceptable rearing tilapias (Popma & Masser 1999; El-Sayed 2006;
when it is not higher than 2 (Craig 2009) but the accept- Mjoun et al. 2010). However, there has been no compre-
able level can vary with the feed price. Feed cost is the hensive analysis of the actual performance of Nile tilapia in
major cost in fish farming (El-Sayed 1999; Craig 2009) farm systems that provide the critical information as to
representing over 50% of the variable costs during the how best to address the yield gap for this globally impor-
grow-out period (El-Sayed 1999). In places where the tant aquaculture resource – either through improved hus-
feed price is high, a small increase in FCR could consid- bandry or through selective breeding. The objective of this
erably increase the variable cost. Therefore, underperfor- study was to quantify the effects of the most likely environ-
mance in terms of FCR is a major concern for mental and management factors on FCR, mortality and
aquaculture as it strongly and negatively affects the prof- growth of Nile tilapia and to identify the most important of
itability of fish farms. these factors associated with the yield gap.
Both primary determinants of FCR at the production
unit level, mortality and individual differences between fish
Material and methods
in converting feed to biomass, are strongly influenced by
the environment (de Verdal et al. 2018). Mortality, espe- Literature search
cially late mortality, is an important determinant of FCR. A systematic literature search was conducted for peer-
Rates of mortality for Nile tilapia vary considerably, with reviewed journal articles that had been published in English
20–71% mortality being reported for Nile tilapia reared in in the ASFA (1971–2016) and CAB-Abstracts (1979–2016)
fertilized ponds with or without supplementary feeding databases on the 7th of July 2016. We used the following
(Abdalla et al. 1996; Abdelghany & Ahmad 2002). Accord- search terms and Boolean operators (‘feed efficiency’ OR
ing to Rana and Hassan (2013), the reported mortality var- FCE OR ‘feed conversion’ OR FCR OR ‘growth rate’
ies between 25 and 60% in pond environments. Trọng OR survival OR mortality) AND (‘Nile tilapia’ OR

Reviews in Aquaculture (2020) 12, 524–541

© 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd 525
S. B. Mengistu et al.

‘Oreochromis niloticus’). Based on the above search terms, were extracted from each article. In most of the studies, the
we found 889 and 1739 articles from ASFA and CAB- numbers of fish used in the experiments or the standard
Abstract databases respectively. The two searches were errors were not reported and thus we gave equal weight to
combined and duplicates were removed using EndNoteX7. all the studies.
This resulted in 1973 articles, which were then checked From the extracted variables CP, water temperature, pH
against the search terms in the title and abstract, which and DO are environmental variables while the rest are man-
resulted in 140 eligible peer-reviewed articles. From these agement variables. FCR, survival and growth rate are the
potentially relevant studies, 108 studies were excluded for key determinants of productivity. To allow for comparisons
one of the following reasons: (i) because articles were not across studies on growth rate, we calculated the thermal
pffiffiffiffiffiffi pffiffiffiffiffiffiffi
accessible (21 studies), (ii) because they did not report a growth coefficient (TGC) as ½ð 3 Wt
3 W0 Þ=ðT  t Þ
sufficient proportion of the variables included in the differ- 1000 where Wt and Wo are final and initial weights, respec-
ent models (20 studies), or (iii) because studies were out- tively, T is the average temperature during the growth per-
side the scope of this review. Studies on the effect of density iod and t is the length of the growth period (Jobling, 2003).
on survival during transportation, lethal dose of salinity, Therefore, the key traits analysed in this study were FCR,
compensatory growth with feed restriction and refeeding, survival and TGC.
sex reversal, or varying crude protein levels during the We first did a principal component analysis (PCA) using
study period were considered as being outside the scope of prcomp package in R software (R Core Team, 2015) to
the review (Fig. 1). The data were extracted from the explore the explanatory variables. If variables were missing
remaining studies for analysis. for some studies, we used the mean values for those vari-
ables and used all the 32 studies in the PCA. Next, we per-
formed linear mixed models to estimate the effects of the
Data extraction and statistical analysis
explanatory variables on FCR, survival and TGC. The
We extracted data on the following variables: study (since explanatory variables were study, study length, stocking
each study can be regarded as a separate element), ‘study density, SW, CP levels, DO, temperature, pH, feeding rate,
length’, which is the grow-out period studied, stocking feeding frequency, the quadratic terms of CP levels, DO,
density, feeding rate, feeding frequency, levels of crude pro- temperature and pH. Only a few studies reported salinity,
tein (CP) in the diet expressed as percentage, stocking ammonia, nitrate and nitrite and therefore the effects of
weight (SW), which is the weight at the beginning of the these variables were not investigated. Linear mixed models
experiments, harvest weight (HW), water temperature, pH, were used to account for the variation in studies and study
dissolved oxygen (DO), salinity, ammonia, nitrate, nitrite, was fitted as a random variable, whereas the rest were fitted
growth and survival. We also extracted FCR or calculated it as fixed effects. All models were analysed using the lme4
as the inverse of total biomass harvested/total feed given. package (Bates et al. 2015) for R software (R Core Team,
Based on the number of treatments within experiments in 2015). The significance of fixed effects was based on the
an article, multiple data records or results of treatments approximate Student’s t-test (Bates et al. 2015). The

Number of search Number of search results

results ASFA 889 CAB-Abstract 1738

Number of papers excluded based

on multiple occurrences 654
Number of potentially relevant papers
after removal of duplicates 1973
Number of papers excluded based
on title/abstract 1833
Number of potentially relevant papers
based on title/abstract 140

Number of papers excluded with

reasons 108
Number of included papers 32

Figure 1 Flow diagram of article selection process.

Reviews in Aquaculture (2020) 12, 524–541

526 © 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd
 Determinants of yield gap in Nile tilapia farms

non-significant effects were removed stepwise, leaving out FCR ¼ b0 þ b1  D þ b2  SW þ b3  CP þ b4  DO

the factor with the highest P-value. þ b5  T þ b6  Feeding rate þ Study þ e
ð3:1Þ
FCR
The majority of papers reported DO, but pH, feeding rate
In all of three models, FCR equals feed conversion
and feeding frequency were not reported in all the studies.
ratio, b0 is the overall intercept, b1 to b10 are the
Hence three separate analyses were undertaken, each with a
regression coefficients of the different explanatory vari-
different model.
ables on FCR, Study is a random study effect assumed
Model 1 was used for studies that reported study length
to be normally distributed (Nð0; r2study ), ɛ is a residual
(L), stocking density (D), SW, CP, DO and temperature
random error assumed to be normally distributed
(T). The final analysis was based on 179 data records from
(Nð0; r2e ), r2study is the variance due to study and r2e is
28 studies that report FCR:
the residual variance.
FCR ¼ b0 þ b1  L þ b2  D þ b3  SW þ b4  CP þ b5
Survival
 DO þ b6  T þ b7  CP2 þ b8  DO2 þ b9
Model 4 was used to investigate the effect of study length,
 T 2 þ Study þ e
stocking density, SW, CP, DO and temperature on survival,
ð1Þ based on 187 data records from 29 studies:
After removing the non-significant effects, the reduced
Survival ¼ b0 þ b1  L þ b2  D þ b3  SW þ b4  CP
model was:
þ b5  DO þ b6  T þ b7  CP2 þ b8  DO2
þ b9  T 2 þ Study þ e
FCR ¼ b0 þ b1  SW þ b2  CP þ b3  DO þ b4  T
þ b5  T 2 þ Study þ e ð4Þ
ð1:1Þ
The effects of CP, DO and temperature were not signifi-
Model 2 used a subset of 23 studies out of the 27 used in cant which led to the following reduced model:
model 1 that also reported pH which resulted in 141 data
records: Survival ¼ b0 þ b1  SW þ Study þ e ð4:1Þ

FCR ¼ b0 þ b1  L þ b2  D þ b3  SW þ b4  CP þ b5 b0 is the overall intercept, b1 is the regression coefficient

of SW on survival, Study is a random study effect assumed
 DO þ b6  T þ b7  pH þ b8  CP2 þ b9
to be normally distributed (Nð0; r2study ), ɛ is a residual ran-
 DO2 þ b10  T 2 þ b11  pH 2 þ Study þ e
dom error assumed to be normally distributed (Nð0; r2e ),
ð2Þ r2study is the variance due to study and r2e is the residual
variance.
After removing the non-significant effects, the reduced
A few studies on survival reported pH, feeding rate and
model was:
feeding frequency, hence a separate set of analyses was done
FCR ¼ b0 þ b1  CP þ b2  DO þ b3  T þ b4  pH to investigate the effect of these explanatory variables, but
none of them were significant and details of these models
þ b5  T 2 þ Study þ e
are not presented here.
ð2:1Þ

Model 3 used a second subset of 11 studies out of the 27 TGC

used in model 1 that also reported feeding rate and feeding Model 5 was used to investigate the effect of study length,
frequency which resulted in 67 data records: stocking density, SW, CP and DO on TGC. This model was
fitted on 192 data records from 29 studies that reported
FCR ¼ b0 þ b1  L þ b2  D þ b3  SW þ b4  CP þ b5 TGC:
 DO þ b6  T þ b7  Feeding rate þ b8
 feeding freq. þ b9  CP2 þ b10  DO2 þ b11 TGC ¼ b0 þ b1  L þ b2  D þ b3  SW þ b4  CP þ b5
 T 2 þ Study þ e  DO þ b6  CP2 þ b7  DO2 þ Study þ e

ð3Þ ð5Þ

After removing the non-significant effects, the reduced After removing the non-significant effects, the reduced
model was: model was:

Reviews in Aquaculture (2020) 12, 524–541

© 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd 527
S. B. Mengistu et al.

TGC ¼ b0 þ b1  L þ b2  D þ b3  CP þ b4  DO DO, pH and feeding rate were positive. Stocking density

þ Study þ e and temperature were negatively correlated with DO, pH
and feeding rate, whereas SW was negatively correlated
ð5:1Þ
with CP, DO, pH and feeding rate. Study length was nega-
Model 6 was applied to a subset of 23 studies out of the tively correlated with CP, feeding rate and DO (Fig. 2).
29 studies used in model 5 that also reported pH resulted
in 155 data records: Feed conversion ratio

TGC ¼ b0 þ b1  L þ b2  D þ b3  SW þ b4  CP þ b5 The linear effects of CP, DO and temperature on FCR were

 DO þ b6  pH þ b7  CP2 þ b8  DO2 þ b9 significant in all three models (1, 2 and 3, P < 0.05, Fig. 3a),
 pH 2 þ Study þ e whereas the quadratic term of temperature was significant in
model 1 and 2 but not in model 3 when corrected for feeding
ð6Þ
rate (Table 1). The positive quadratic term of temperature
After removing the non-significant effects, the reduced in models 1 and 2 indicated that the relationship between
model was: FCR and temperature was not linear as demonstrated clearly
in Figure 3b. The FCR was above 2.0 when the temperature
TGC ¼ b0 þ b1  D þ b2  CP þ b3  DO þ b4  pH was below 26°C and above 33°C. Optimum FCR was
þ Study þ e between 27°C and 32°C. FCR increased dramatically when
ð6:1Þ the temperature drops below 25°C and reaching 4.4 at 20°C.
Increasing levels of CP (15
50.7%) and DO
Few studies reported feeding rate and feeding frequency (1
11.1 mg/L) decreased FCR in all three models
together with TGC, hence we did a separate set of analyses, (P < 0.05, Table 1, Fig. 3a), as did increasing pH
each with a different model, to investigate the effect of these (6.42
8.3) in model 2 (P < 0.001, Table 1, Fig. 3c). Other
variables on TGC. variables tested in more than one model did not show a
Model 7 used another subset of 14 studies from model 5 consistency of response or were not significant. FCR
that reported feeding rate and feeding frequency in addi- increased significantly with increasing stocking density
tion to the other variables fitted in model 5, which resulted when corrected for feeding rate in model 3 (P = 0.017), but
in 86 data records: was not significant (P > 0.05) in model 1 and 2. The effect
of SW on FCR was significant and positive (0.003,
TGC ¼ b0 þ b1  L þ b2  D þ b3  SW þ b4  CP þ b5 P < 0.001) in model 1, not significant in model 2
 DO þ b6  feeding rate þ b7  Feeding freq. (P = 0.084) but significant and negative (
0.016,
þ b8  CP2 þ b9  DO2 þ Study þ e P = 0.017, Table 1) when corrected for feeding rate in
ð7Þ model 3. In model 1 and 2 the SW range is similar while in
model 3 it is much smaller. The difference in the sign of
After removing the non-significant effects, the reduced coefficients of SW in model 1 and 2 is most likely due to the
model was: difference in SW ranges in the two models (0.003–311 g
and 0.012–110 g, Table 1). The effect of feeding frequency,
TGC ¼ b0 þ b1  feedingrate þ Study þ e ð7:1Þ the quadratic terms of CP levels, DO and pH on FCR were
not significant (P > 0.05) in any of the three models.
With TGC being thermal growth coefficient, b0 is the In summary, FCR decreased with increasing CP, DO and
overall intercept, b1 to b9 are the regression coefficients of pH, and was optimal in a temperature range from 27.0–
the different variables on TGC, Study is a random effect 32.0°C. Results were inconsistent for stocking density and
assumed to be normally distributed (Nð0; r2study ), ɛ is a SW. Among the environmental variables, temperature had
residual random error assumed to be normally distributed the largest effect on FCR followed by DO and pH.
(Nð0; r2e ), r2study is the variance due to study and r2e is the
residual variance.
Survival
The studies used in each model are given in Appendix 1.
The analysis of model 4 showed a significant effect of only
SW (P = 0.025, the linear equation is: Survival ¼
Results
89:767 þ 0:03  SW) on survival and no significant effect
Principal component analysis of any of the other variables (P > 0.05). Survival increased
The first two principal components explained 42% of the by 0.03% per gram increase in SW. Increasing stocking
variation in the whole data set. The correlations among weight from 5 to 50 g would improve survival by 1.4%

Reviews in Aquaculture (2020) 12, 524–541

528 © 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd
 Determinants of yield gap in Nile tilapia farms

Figure 2 Loading plot from principal component analysis of all the data points from 32 studies.

(Fig. 4). The effects of feeding frequency, feeding rate and the environmental variables DO had the largest effect on
pH on survival were not significant for the range of values TGC while, as expected, feeding rate had the largest effect on
investigated (results not shown). TGC from the management variables investigated.

Thermal growth coefficient (TGC) Discussion

TGC increased with increasing levels of CP (15
50.7%) The main environmental and management factors influenc-
and DO (1
11.1 mg/L) tested in models 5 and 6 ing survival, FCR and growth of Nile tilapia in the 32 papers
(P < 0.05, Tables 2, Fig. 5a) but when corrected for feeding identified in a systematic literature survey were DO, temper-
rate in model 7 these effects became not significant. TGC ature, pH, CP, SW, feeding rate and stocking density.
increased with increasing pH (6.42
8.2) in model 6 Ammonia, nitrite, nitrate and salinity are important water
(P = 0.001, Table 2, Fig. 5b) and feeding rate (2
60%) in quality parameters worth of inclusion in the analysis but
model 7 (P = 0.030, the linear equation is: TGC ¼ 0:611 data on these parameters were only available in few studies
þ0:01  feeding rate). and therefore these parameters were not investigated. We
Increased stocking density decreased TGC significantly in discussed the main environmental and management factors
models 5 and 6 (P < 0.05) but not in model 7 (P > 0.05) influencing yield gap focusing mainly on pond production
which included feeding rate. The effect of study length on which is the predominant production system. The PCA
TGC was significant in model 5 (P < 0.001), but not in analysis showed a correlation between explanatory variables.
models 6 and 7 (P > 0.05). The effect of feeding frequency Pearson correlations between the explanatory variables were
and the quadratic term of CP levels on TGC were not signifi- non-significant to weak or moderate correlations. The high-
cant (P > 0.05). In summary, TGC increased with increasing est correlation between stocking density and stocking weight
CP, DO and pH and decreased with increasing stocking den- was 0.57 (P < 0.01). Using median values of the significant
sity and study length, although not in every analysis. Among variables and coefficients from Table 1, model 1 and varying

Reviews in Aquaculture (2020) 12, 524–541

© 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd 529
S. B. Mengistu et al.

(a) (b)

(c)

Figure 3 (a) The effect of dissolved oxygen on FCR calculated based on the coefficient estimates from Table 1, model 1 and median values
(SW=9.3, CP=34% and Temp. = 28°C) for other variables. The linear equation is: FCR ¼ 32:4 þ 0:003  9:3
0:029  34
0:102  DO

1:99  28 þ 0:034  282 , Modelled, Raw data. (b) The effect of temperature on FCR calculated based on the coefficient estimates from
Table 1, model 1 and median values (SW=9.3, CP=34%, and DO=6.05) for other variables. The linear equation is: FCR ¼ 32:4 þ 0:003
9:3
0:029  34
0:102  6:05
1:99  Temp: þ 0:034  Temp:2 , Modelled Raw data. (c) The effect of pH on FCR calculated based on
the coefficient estimates from Table 1, model 2 and median values (CP=31%, DO=6.05 and Temp. = 28°C) for other variables. The linear equation is:
FCR ¼ 38:615
0:034  31
0:101  6:05
2:107  28
0:579  pH þ 0:036  282 , Modelled Raw data.

Reviews in Aquaculture (2020) 12, 524–541

530 © 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd
 Determinants of yield gap in Nile tilapia farms

Coefficient  SE

The number of studies and data records (in parentheses) utilized in each model are given below the model number and the studies are listed in detail in Appendix 1. Significance levels are indicated as
1.5***

0.0***

0.1***

 0.0***
0.0*
0.0*

0.1*







0.050
0.337
8.7
0.1

0.0

0.0

0.2

0.2

0.1
NA
Parameter range

0.012–110.2
0.003–22.0

17.2–50.7
3.2–10.9
20–31.5
Model 3
11 (67)†

2–60
NA
Table 1 Regression coefficient estimates  standard errors given to one decimal place for factors that affect FCR for reduced Models 1, 2 and 3.

Coefficient  SE

39.0  4.3***

0.0***

0.3***
0.2***

0.0***
0.0**

0.0 

0.1 

2.2 

0.6 

0.0 
0.118
0.289
NA
Parameter Range

6.42
8.3
23 (141)†

Figure 4 The effect of stocking weight on survival. The fitted line was
15–50.7
Model 2

1–10.9
20
34

based on the estimated coefficients from model 5 and varying stocking

NA

weight from 4 to 50 g. The resulting equation is: Survival ¼ 89:767þ

0:03  stocking weight, Modelled Raw data.
Coefficient  SE

32.4  3.5***

0.0***

0.3***

0.0***
0.0**
0.0**

DO from the lowest values to the highest value improved

FCR by 50%. Using median values of the significant vari-
0.0 

0.0 

0.1 

2.0 

0.0 
0.112
0.248

ables and coefficients from Table 1, model 1 and varying

NA
NA

temperature from 20 to 29.5°C improved FCR by 68%.

Using median values of the significant variables and coeffi-
cients from Table 1, model 2 and varying pH from the low-
est values to the highest value improved FCR by 46%. Using
Parameter range

the median values of the significant variables from Table 2,

0.012
311.1

model 5 for DO and from Table 2, model 6 for pH and

28 (179)†

15–50.7
Model 1

varying DO and pH levels from minimum to maximum

1–11.1
20–34

improved TGC by 88 and 52% respectively (Fig. 5a,b).

NA
NA

These results are now discussed with a view to determine

whether changes to husbandry practices can reduce the yield
*P < 0.05, **P < 0.01, ***P < 0.001.

gap or whether it is possible to provide solutions through

selective breeding for those variables that are difficult or
Feeding rate (% of body weight)

impossible to control in given farming systems.

Stocking density (kg m
3)

Tilapia farmers practice and the effects of husbandry

Temperature (°C℃)

Residual variance

management
Study variance
Temperature2
Parameters

DO (mg/L)

Stocking weight and study length

Intercept

CP (%)
SW (g)

In this study, we found significant effects of stocking weight

pH

(SW) on FCR, when corrected for feeding rate, and on

†

Reviews in Aquaculture (2020) 12, 524–541

© 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd 531
S. B. Mengistu et al.

Table 2 Regression coefficient estimates  standard errors given to one decimal place for factors that affect TGC for Models 5 and 6.

Model 5 Model 6
29 (192)† 24 (155)†
Parameters Parameter range Coefficient  SE Parameter range Coefficient  SE

Intercept 0.4  0.2*
1.1  0.4*

Study length (days) 25–196
0.0  0.0*
Stocking density (kg m
3) 0.003–41.4
0.0  0.0** 0.003–39.0
0.0  0.0*
CP (%) 15–50.7 0.0  0.0** 15–50.7 0.0  0.0*
DO (mg/L) 1–11.1 0.1  0.0** 1–11.1 0.1  0.0**
pH NA NA 6.42–8.2 0.2  0.1***
Study variance 0.158 0.200
Residual variance 0.027 0.024
†
The number of studies and data records (in parentheses) utilized in each model are given below the model number and the studies are listed in detail
in Appendix 1. Significance levels are indicated as *P < 0.05, **P < 0.01, ***P < 0.001.

(a) (b)

Figure 5 (a) The effect of dissolved oxygen on TGC. The modelled data were calculated based on the coefficient estimates from Table 2, model 5 and
median values (study length = 70 days, stocking density = 0.894 and CP = 34%) for other variables. The resulting linear equation is:
TGC ¼ 0:436
0:003  70
0:014  0:894 þ 0:010  34 þ 0:047  DO, Modelled Raw data. (b) The effect of pH on TGC. The modelled data
were calculated based on the coefficient estimates from Table 2, model 6 and median values (stocking density = 0.894 and CP=34%) for other variables,
resulting in the following equation: TGC ¼
1:128
0:011  0:894 þ 0:01  34 þ 0:047  DO þ 0:191  pH

survival, with survival (Fig. 4) and FCR (Fig. 6) increasing dependent cannibalism for Nile tilapia weighing 0.03 to
with increasing SW. It is clearly seen from Figure 4 that 15.08 g. They found a sigmoid relationship between preda-
increasing SW increased survival in the range of SW 4 to tor to prey weight ratio and the probability of prey being
10 g, whereas the relationship looks like a sigmoid curve killed. This would explain the sigmoid relationship between
when considering the whole range from 4 to 50 g. Fesse- SW and survival. Stocking fish larger than 10 g and graded
haye et al. (2006) found significant mortality due to size for size uniformity could help to avoid size dependent

Reviews in Aquaculture (2020) 12, 524–541

532 © 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd
 Determinants of yield gap in Nile tilapia farms

Figure 6 The effect of SW and dissolved oxygen on FCR using coeffi- Figure 7 The effect of stocking density on FCR in low oxygen (3 mg/l)
cients from Table 1, model 1 and the median values (CP=34%, and high oxygen (5 mg/l) levels. FCR was calculated based on the coef-
T = 28°C) for the variables while varying SW from 5 to 60 g and fixing ficient estimates from Table 1, model 3 and varying the density level for
dissolved oxygen (DO) at 3mgl-1(hypoxia) or 5mgl-1 (normoxia) 5 g and 10 g fish, fixing dissolved oxygen level to 3 or 5 mg/l and med-
(FCR ¼ 32:4 þ 0:003  SW
0:029  34
0:102  DO
1:999  T þ ian values (stocking density = 0.06, CP=34%, Temp. = 26°C and feed-
0:034  T 2 Þ, DO 3 mg/l 5 mg/l. ing rate = 4% of body weight) for the other variables (FCR ¼ 8:728
þ0:097  0:06
0:016  SW
0:048  34
0:165  DO
0:163
26 þ 0:047  4Þ, DO.SW ● DO 3 mg/l, SW 10 g ▲ DO 3 mg/l, SW 5 g
cannibalism at smaller SW. The ranges of stocking density ■ DO 5 ml/l, SW 10 g + DO 5 mg/l, SW 5 g.
tested in our models were from 0.003 to ~22–41 kgm
3.
When keeping DO constant at 3 mg/L or 5 mg/L and vary-
ing stocking density from 1
20 fish per cubic meter, FCR removed, the effect of study length on TGC turned from
and TGC hardly changed indicated by the almost flat lines statistically significant to non-significant. These studies are
(Figs 7 and 8). However, when keeping stocking density highly influential because the study length is relatively short
constant and increasing DO from 3 to 5 mg/L, FCR at 25–30 days (average study length was 87.44 days) and
reduced from 2.3 and 2.4 to 2.0 and 2.1 and TGC increased the studies contributed 19 data points to the analysis.
from 0.77 to 0.88 (Figs 7 and 8). Stocking density, corrected for feeding rate, had a signifi-
Under small-holder tilapia farm conditions diurnal DO cant effect on FCR (model 3), but it was not significant in
fluctuation is very high. Therefore, stocking densities of 3– model 1 and 2. It also had a significant effect on TGC
5 fish of size larger than 10 g per square meter would give a (model 5 and 6). Increasing stocking density negatively
better result than stocking smaller and/or more fish (Figs 6, affected both FCR and TGC. This agrees with what is gen-
7 and 8). erally observed in aquaculture (Ellis et al. 2002; Papout-
The effect of study length on FCR was not significant soglou et al. 2006; Li et al. 2012). Our estimates of the
while it was significant on TGC in model 5 while not signif- regression of stocking density on FCR (0.097) and stocking
icant in models 6 and 7. This is due to the fact that the density on TGC (
0.014) suggest that increasing stocking
analysis with model 5 has more data points with short density by one unit would lead to an increase in FCR by
study length. When two studies with short study length about 0.01 kg feed per kg biomass harvest and a reduction
(El-Sayed & Teshima 1992; Tran-Duy et al. 2008) were in TGC by 0.014.

Reviews in Aquaculture (2020) 12, 524–541

© 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd 533
S. B. Mengistu et al.

stocking density in Thailand is two to five fish (Bhujel

2013). The effect of stocking density is dependent on DO
levels. Figs 7 and 8 suggest that 3–5 fish would give better
FCR and TGC in a low oxygen environment. The above
stocking densities used in different countries are in agree-
ment with this study and will result in good FCR and TGC.

Feeds and feeding

FCR (models 1, 2 and 3) and TGC (models 5 and 6)
improved with increasing CP. Model 3 and model 7 showed
that both FCR and TGC increased with increasing feeding
rate. The effect of feeding rate on FCR and TGC is well
described in the literature (e.g. review by de Verdal et al.
2018). As in terrestrial animals, protein plays a vital role in
fish. It constitutes about 65–75% of fish body weight on dry
matter basis (Halver & Hardy 2002). Fish require protein for
growth, development and reproduction. Protein deficient
feeds can negatively affect growth or lead to interruption of
growth and loss of weight (Halver & Hardy 2002). Feed cost
constitutes the major portion of the variable cost in fish
farming (El-Sayed 1999) and protein is the most expensive
feed ingredient. Profitability is a key factor in any commercial
fish farming system. Therefore cost effective feed composition
that can satisfy nutritional requirements and feed manage-
ment that can optimize FCR and TGC is crucial.
The CP requirement for starter, grower and finisher fish
Figure 8 The effect of stocking density on TGC in low oxygen (3 mg/l) is 30–35%, 30–32% and 28–30% respectively (FAO, 2018).
and high oxygen (5 mg/l) levels. TGC was calculated based on the coef- Least-cost feed contains 20, 25 and 30 CP levels for finisher,
ficient estimates from Table 2, model 5 and varying the density level for
grower and starter, respectively, at a feeding rate of 1.5–5%
5 g and 10 g fish, fixing dissolved oxygen level to 3 or 5 mg/l and med-
ian values (study length = 70, CP=34%) for the other variables
body weight and 3–4 feeds per day (Ng & Romano 2013).
(TGC ¼ 0:436
0:003  70
0:014  D þ 0:011  34 þ 0:56  DOÞ, El-Saidy and Gaber (2005) found that the economic opti-
DO. SW ● DO 3 mg/l, SW 10 g ▲ DO 3 mg/l, SW 5 g ■ DO 3 mg/l, mum is at 25% CP and feeding rate of 2% compared with
SW 10 g + DO 5 mg/l, SW 10 g DO 5 mg/l, SW 5 g. 30% CP and 2% feeding rate for adult Nile tilapia reared in
concrete tanks. According to a review by El-Sayed (2013),
most smallholder farmers in sub-Saharan Africa fertilize
Under small-scale tilapia production systems, stocking their ponds to boost natural feed. In addition, some farm-
density, number of fish per square meter and stocking size ers use farm-made feed, cotton seed cake, wheat bran, rice
differ from country to country. In Malaysia, five fish of 5 g bran or maize bran for supplementary feeding. In Thailand
per square meter are stocked (Azlan Bin Azizan, pers. and the Philippines, small-holder tilapia farmers fertilize
comm., 2017); in China, 4–6 fish of 4 g on average are their ponds and use commercial feed, cereal brans, restau-
stocked per square meter in summertime, while in the win- rant wastes or bakery wastes as supplementary feeding
ter time they stock bigger fish, on average 18 g (Liu et al. (Bhujel 2013; Romana-Eguiaa et al. 2013). According to a
2013). In the Philippines, stocking density in extensive pro- review by Rana and Hassan (2013), the CP content of tila-
duction systems is less than one fish of 10–20 g per square pia feed used ranges between 16 and 32%. CP and feeding
meter, in semi-intensive systems it is 1–5 and in intensive rate can be easily managed to optimize production and
systems it is 5–10 fish of the same size, but in intensive sys- should be kept at the optimum level to maximize profit
tems using aeration, the preferred stocking size by farmers and minimize yield gap. The optimum feeding rate is the
is five to twenty grams (Romana-Eguiaa et al. 2013). The rate that gives the lowest FCR, this feeding rate is lower
growth period varies from two to nine months depending than the feeding rate required for maximum growth (for
on the targeted market niche (Rana & Hassan 2013, Hoong instance in salmon and trout Lovell 1989). With respect to
Yip Yee, pers. comm., 2016) and therefore the length of the CP, in olive flounder, Kim et al. (2002) found that growth
growth period, that is study length in this study, is not so increased with increasing CP levels up to 55% and then
much determined by its effect on FCR and TGC. The decreased with further increase in CP. This would suggest a

Reviews in Aquaculture (2020) 12, 524–541

534 © 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd
 Determinants of yield gap in Nile tilapia farms

non-linear relationship between growth and CP. Therefore, From 1 to 1.9) and to a reduction in 0.4 unit TGC. Using
we would expect also a non-linear relationship between median values of the significant variables and coefficients
FCR and CP. However, we found a linear relationship from Table 1, model 2 and varying DO from 1 to
between CP and FCR within the range of values tested. 10.88 mg/L improved FCR by 50% (Figs 3a). Using coeffi-
Most of the studies used in this systematic review may have cients and the median values of the significant variables
used CP levels close to the optimal or lower than the opti- from Table 2, model 5 and varying DO from 1 to
mal levels with respect to FCR. If feed cost is increasing 11.05 mg/L improved TGC by 88% (Figs 5a). The effect of
with CP, the economic optimum CP would be even lower DO on FCR is larger than the effect of pH, but lower than
than the CP that results in minimum FCR. Feed cost is the the effect of temperature, whereas the effect of DO on TGC
major cost in fish farming (El-Sayed 1999; Craig 2009) and is larger than the effect of pH.
among the feed ingredients CP is the most expensive. DO is one of the main limiting environmental variables
Therefore, feeding should be optimized to the level where that affect fish performance. Low DO affects feed intake
marginal feed cost is equal to marginal revenues and where negatively (Wang et al. 2009) and reduces digestibility
yield gap is minimal. (Tran-Duyn et al. 2012). At high DO, feed assimilation is
The effects of farm-made feeds, supplementary feeds and improved, which may be due to improved blood flow to
pond fertilization on yield gap were not investigated in our the gastrointestinal tract (Axelsson et al. 2002) and lower
analysis. Farm-made feeds may vary in their nutrient con- energy cost of feed digestion and absorption of nutrients
tent depending on the ingredients used. The CP content, (Duan et al. 2011). Therefore, more energy is available for
CP source (NRC, 2011) and CP to energy ratio can affect growth. Tran et al. (2016) found Nile tilapia performed sig-
feed efficiency (Kabir et al. 2019). Algae is a natural feed nificantly less in terms of final body weight, specific growth
source for tilapia and the contribution of algae to tilapia rate and FCR under hypoxia (3mgL
1 ) compared with
growth is estimated to be between 40 and 68% in small- under normoxia 5 mg/L which is 50% of saturation. They
holder tilapia farms (Kabir et al. 2019). The amount of also found that hypoxia affected intestinal morphology
pond fertilization affects algae production in the pond. To negatively. Therefore, optimum DO is a very important
reduce yield gap feeds should contain the optimum amount environmental factor for improving FCR and TGC.
of nutrients and Bhujel (2014) recommends to maintain In non-aerated ponds, DO levels fluctuate during the day
Secchi disc visibility at 30–40 cm depth for appropriate and will be somewhere 0 – 15 mg/L with the highest values
amounts of algae. If future studies include information on in the afternoon and the lowest values just before sunrise
the type of farm made feeds, CP contents, CP source, CP to (Bhujel 2014). However, DO level should be kept at least
energy ratio and Secchi disc visibility, future meta-analyses 5 mg/L and when it drops to ≤ 3 mg/L, feeding should be
could include these parameters to quantify the contribution stopped and remedial action should be taken to improve the
of these factors to the yield gap, which would help in fur- DO levels (Stickney 2017). Pond aeration keeps DO at an
ther minimizing the yield gap. acceptable level with minimal fluctuations. However, DO is
often beyond control in many small-scale farms where aera-
tion for fishponds is not available or too expensive.
Environmental factors
In areas where aeration is available, ponds should be aer-
Dissolved oxygen ated during critical times of the day especially early in the
In this study, we found significant effects of DO on FCR morning and on cloudy days. Managing the algae load in
(models 1, 2 and 3) and on TGC (models 5 and 6) with the water to optimal levels also helps in minimizing the DO
FCR and TGC improving with increasing DO. The effect of demand during the night and prevents a large drop of DO.
the quadratic term of DO on FCR was not significant. Here Usually DO is not a problem in flowing rivers due to ample
we found only a significant linear association between DO water movement, in lakes it can become a problem when it
and FCR, whereas the relationship might actually be curvi- is highly eutrophic which results in algae bloom and
linear since there will be a DO level beyond which FCR will hypoxia during nights. If aeration of ponds is not possible,
no longer improve. The reason that we did not find a curvi- it is clear that there is a need for fish that are resilient to
linear relation might be due to a lack of data points in the low DO levels during parts of the day with low FCR and
lower concentration range. Interestingly, DO had no signif- high TGC despite the extreme DO variation.
icant effect on survival, at least not in the studies that were
analysed in this paper. Our estimate of the regression of Temperature
DO on FCR (-0.111) and DO on TGC (0.056) suggests that Temperature had a significant effect on FCR (models 1, 2
decreasing DO from the highest level investigated 11 mg/L and 3), while it had no significant effect on survival (model
to 3 mgL, which is the minimum level required for tilapia 4). The significant positive quadratic term clearly showed
production, would lead to an increase in 0.9 unit FCR (e.g. that the relationship between FCR and temperature is non-

Reviews in Aquaculture (2020) 12, 524–541

© 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd 535
S. B. Mengistu et al.

linear. FCR was optimum between 27 and 32.0°C and

Implications for management and breeding
increased significantly when the temperature dropped
below 25°C reaching 4.4 at 20°C (Fig. 3b). Nile tilapia per- We conclude that CP, DO, water temperature, pH, stocking
forms best in the upper end of the optimal temperature density and feeding rate are the most important variables
range of 27–32°C, which is in agreement with older reports to take into account to reduce the yield gap in tilapia farm-
quoting 29–31°C being the optimal temperature range for ing. Ammonia, nitrite, nitrate, salinity and Secchi disc visi-
Nile tilapia (Popma & Lovshin 1996; Popma & Masser bility are important water quality parameters but they were
1999). Note that the quadratic term of temperature was not not investigated due to very few studies reporting these
significant in model 3 when accounting for feeding rate, parameters. However, optimizing DO, pH, stocking density
which is most likely due to the fact that feeding rate was and feeding rate positively affects ammonia, nitrite, nitrate,
adjusted for temperature in the studies concerned. Using except for salinity. Low DO and high ammonia are not
median values of the variables and coefficients problematic in flowing rivers due to ample water move-
from Table 1, model 1 and increasing temperature from 20 ment. Salinity is a problem in areas with brackish water
to 29.5°C improved FCR by 68%, which was the highest because Nile tilapia is a fresh water fish and less tolerant to
effect compared with DO and pH (Fig. 3b). Increasing salinity compared with other Oreochromis spp. (Watanabe
temperature within the tolerable range increases et al. 1985). Temperature is practically beyond control in
appetite, food consumption rate and accelerates digestion most farms. Tilapia farms should give emphasis to manag-
of feed (Brett & Groves 1979; Jobling 1993). Management of ing optimal stocking density and feeding rate. DO and pH
water temperature in ponds and cages is not practical; thus, are largely influenced by aeration and liming could
optimizing temperature is not possible. Therefore, it can be improve pH when tilapia are grown in ponds. At present
concluded that it is important to select fish under condi- large numbers of small-scale farmers have no means to aer-
tions that are similar to the prevailing temperatures in ate their ponds, either because it is too expensive, or
commercial environments to optimize FCR. because they have no access to cheap electricity. Breeding
programs should consider this. Selection for higher growth
pH rate will increase feed intake and consequently oxygen con-
Our estimates of the regression of pH on FCR (
0.548) sumption (Omasaki et al. 2017). As the selection environ-
and pH on TGC (0.191) suggest that increasing pH by one ment is usually well managed, with optimal conditions in
unit from 6.42 to 7.42 would improve FCR by about 0.5 terms of DO, pH and CP, there is a risk for genotype by
unit and TGC by 0.2 unit, respectively. Using median val- environment interaction (GxE) when improved strains are
ues of the significant variables and coefficients from used in low-input ponds and a yield gap is expected
Table 1, model 2 and varying pH from 6.42 to 8.3 because of lower production than what is genetically possi-
improved FCR by 46%. Using values from Table 2, model ble in an optimum environment.
6 and the same approach as above, increasing pH from 6.42 In the GIFT breeding program, Ponzoni et al. (2011)
to 8.2 improved TGC by 52%. The factors pH and DO had reported a genetic gain of 10–15% per generation for
a comparable effect on FCR, whereas the effect of pH on growth. In the presence of GxE interaction, the same gain
TGC is half of the effect of DO on TGC. In line with our might not be attained in the production environment when
analysis, Popma and Masser (1999) found the best FCR DO and temperature are far from the optimum levels and
and growth in a pH range from 7 to 9. However, un- create a large difference with the selection environment.
ionized ammonia, which is toxic to fish, increases with Estimates of the degree of GxE for growth in Nile tilapia
increasing pH and water temperature (Randall & Tsui between different rearing environments are inconclusive
2002). Therefore, in order to achieve best results pH should (Sae-Lim et al. 2016). Charo-Karisa et al. (2006) found a
be maintained between 7 and 8. This can be practically low genetic correlation ð
0:27  0:69Þ for body weight of
achieved in ponds using lime (Calcium carbonate fry between ponds. Trọng et al. (2013) compared the
(CaCO3)) (Lekang 2013). growth of GIFT Nile tilapia reared in river cages, aerated
Among the environmental factors pH can be easily man- nucleus ponds and non-aerated low-input ponds, and
aged to optimize growth, FCR and survival. Small-scale found a high genetic correlation (0.83) for daily growth
farmers manage water pH using lime, particularly under coefficient (DGC). Eknath et al. (2007) found high genetic
intensive pond production systems, while usually pH is not correlations (0.76
0.99) among different pond environ-
a problem for river cage production systems, where water ments and medium to high genetic correlation (0.36
0.82)
exchange is sufficient to maintain pH at optimum levels. between pond and cage environments. Bentsen et al.
Aeration can also help to reduce the amount of carbon (2012) found high genetic correlations (0.53
0.99,
dioxide that would otherwise interact with water and pro- mean = 0.89) for body weight between different environ-
duce carbonic acid. ments. Robertson (1959) suggested that GxE interactions

Reviews in Aquaculture (2020) 12, 524–541

536 © 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd
 Determinants of yield gap in Nile tilapia farms

are biologically meaningful when the genetic correlation acknowledged. The work was assisted through support pro-
between environments is less than 0.8. GxE interactions vided to WorldFish by the European Commission-IFAD
with genetic correlations between environments of 0.8 or Grant Number 2000001539, the International Fund for
higher are considered not strong. However, if indeed the Agricultural Development (IFAD) and the CGIAR Research
true genetic correlation is 0.8, it means that only 80% of Program on Fish Agrifood Systems (FISH).
the maximum possible genetic gain can be achieved in the
production environment when selection is in the nucleus
References
environment and information of only the selection envi-
ronment is used in genetic evaluations (Mulder & Bijma Abdalla AAF, McNabb CD, Batterson TR (1996) Ammonia
2005). Use of half-sib information from the production Dynamics in Fertilized Fish Ponds Stocked with Nile Tilapia.
environment would reduce the loss in selection response The Progressive Fish-Culturist 58(2): 117–123.
(Brascamp et al. 1985; Mulder & Bijma 2005). Omasaki Abdelghany AE, Ahmad MH (2002) Effects of feeding rates on
et al. (2016a) compared growth of Nile tilapia in a growth and production of Nile tilapia, common carp and sil-
commercial monosex environment and a mixed sex ver carp polycultured in fertilized ponds. Aquaculture
Research 33(6): 415–423.
nucleus environment and found significant GxE (genetic
Abdel-Tawwab M, Hagras AE, Elbaghdady HAM, Monier MN
correlation = 0.59) which was probably caused by the
(2014) Dissolved oxygen level and stocking density effects on
methyl-testosterone treatment to produce monosex fry.
growth, feed utilization, physiology, and innate immunity of
They recommend to use sib information from the monosex
Nile tilapia, Oreochromis niloticus. Journal of Applied Aquacul-
production environment, similar to the general recommen-
ture 26(4): 340–355.
dation by Mulder and Bijma (2005). Lower prices for geno- Abdel-Tawwab M, Hagras AE, Elbaghdady HAM, Monier MN
typing may make it easier to include information of (2015) Effects of dissolved oxygen and fish size on Nile tilapia,
commercial animals in genetic evaluations and reduce the Oreochromis niloticus (L.): growth performance, whole-body
yield gap when using genomic selection (for instance see composition, and innate immunity. Aquaculture International,
Mulder 2016). 23(5), 1261–1274.
Abou Y, Fiogb
e ED, Micha JC (2007) Effects of stocking density
on growth, yield and profitability of farming Nile tilapia, Ore-
Conclusion
ochromis niloticus L., fed Azolla diet, in earthen ponds. Aqua-
We found that temperature had the largest effect on FCR culture Research, 38(6), 595–604.
followed by DO, pH and CP, whereas DO had the largest Al-Hafedh YS (1999) Effects of dietary protein on growth and
effect on TGC followed by CP and pH. Attempting tilapia body composition of Nile tilapia. Oreochromis niloticus L.
farming in regions outside the optimal temperature range Aquaculture Research 30(5): 385–393.
would have a negative effect on production efficiency Alhassan EH, Abarike ED, Ayisi CL (2012) Effects of stocking
unless the strains used are selected for such temperature density on the growth and survival of Oreochromis niloticus
range. Among the management variables, feeding rate had cultured in hapas in a concrete tank. African Journal of Agri-
the largest effect on FCR and TGC followed by stocking cultural Research 7(15): 2405–2411.
density, study length and SW. Management could control Ali A, Al-Ogaily SM, Al-Asgah NA, Goddard JS, Ahmed SI
these variables. Based on this review analysis we recom- (2008) Effect of feeding different protein to energy (P/E)
ratios on the growth performance and body composition of
mend optimizing management in terms of stocking den-
Oreochromis niloticus fingerlings. Journal of Applied Ichthyol-
sity (3
5fishm
2 ), SW (> 10 g), CP (25
30%), DO
ogy 24(1): 31–37.
(> 5 mg/L) and pH (7
8). This will improve FCR, sur-
Axelsson M, Altimiras J, Claireaux G (2002) Post-prandial blood
vival and growth rate and reduce the yield gap in tilapia
flow to the gastrointestinal tract is not compromised during
farming. Temperature has a very large effect on FCR, but
hypoxia in the sea bass Dicentrarchus labrax. Journal of Experi-
it is hard to influence water temperature. DO is largely mental Biology 205(18): 2891–2896.
influenced by aeration when tilapia are grown in ponds. Azaza MS, Dhra€ıef MN, Kra€ıem MM (2008) Effects of water tem-
Since many small and medium sized farms do not have perature on growth and sex ratio of juvenile Nile tilapia Ore-
aeration, these major tilapia farming systems could benefit ochromis niloticus (Linnaeus) reared in geothermal waters in
from genetically improved strains selected for resilience to southern Tunisia. Journal of Thermal Biology 33(2): 98–105.
highly fluctuating diurnal temperature and DO levels. Azaza MS, Assad A, Maghrbi W, El-Cafsi M (2013) The effects
of rearing density on growth, size heterogeneity and inter-
individual variation of feed intake in monosex male Nile tila-
Acknowledgements
pia Oreochromis niloticus L. Animal 7(11): 1865–1874.
This work is part of the first author’s sandwich PhD study, Azaza MS, Khiari N, Dhraief MN, Aloui N, Kra€ıem MM, Elfeki
funded by Koepon Foundation whose support is A (2015) Growth performance, oxidative stress indices and

Reviews in Aquaculture (2020) 12, 524–541

© 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd 537
S. B. Mengistu et al.

hepatic carbohydrate metabolic enzymes activities of juvenile Ellis T, North B, Scott AP, Bromage NR, Porter M, Gadd D
Nile tilapia, Oreochromis niloticus L., in response to dietary (2002) The relationships between stocking density and welfare
starch to protein ratios. Aquaculture Research, 46(1), 14–27. in farmed rainbow trout. Journal of Fish Biology 61(3): 493–
Azevedo R. V. D., Oliveira K. F. D., Flores-Lopes F., Teixeira- 531.
Lanna E. A., Takishita S. S., Tavares-Braga L. G. (2015) El-Saidy DMSD, Gaber MMA (2005) Effect of dietary protein
Responses of Nile tilapia to different levels of water salinity. levels and feeding rates on growth performance, production
Latin American Journal of Aquatic Research, 43(5), 828–835. traits and body composition of Nile tilapia, Oreochromis
Bahnasawy MH (2009) Effect of dietary protein levels on growth niloticus (L.) cultured in concrete tanks. Aquaculture Research,
performance and body composition of monosex Nile tilapia, 36(2), 163–171.
Oreochromis niloticus L. Reared in fertilized tanks. Pakistan El-Sayed AFM (1999) Alternative dietary protein sources for
Journal of Nutrition, 8(5), 674–678. farmed tilapia. Oreochromis spp. Aquaculture 179(1–4): 149–
Bates D, Machler M, Bolker BM, Walker SC (2015) Fitting linear 168.
mixed-effects models using lme4. Journal of Statistical Soft- El-Sayed AFM (2006) Tilapia Culture. CABI, UK.
ware, 67(1): 1–48. El-Sayed AFM (2013) Tilapia feed management practices in sub-
Bentsen HB, Gjerde B, Nguyen NH, Rye M, Ponzoni RW, Palada Saharan Africa. In Hasan MR and New MB (eds) On-farm
de Vera MS et al. (2012) Genetic improvement of farmed tila- feeding and feed management in aquaculture. FAO Fisheries
pias: genetic parameters for body weight at harvest in Nile and Aquaculture Technical Paper No. 583. pp. 377–405. FAO,
tilapia (Oreochromis niloticus) during five generations of test- Rome.
ing in multiple environments. Aquaculture 338–341: 56–65. El-Sayed AFM, Teshima SI (1992) Protein and energy require-
Bhujel RC (2013) On-farm feed management practices for Nile ments of Nile tilapia, Oreochromis niloticus, fry. Aquaculture
tilapia (Oreochromis niloticus) in Thailand. On-farm feed and 103(1): 55–63.
feed management in aquaculture. FAO Fisheries and Aquacul- El-Sherif MS, El-Feky AMI (2009) Performance of Nile tilapia
ture Technical Paper No. 583. pp. 159-189. FAO, Rome. (Oreochromis niloticus) fingerlings. II. Influence of different
Bhujel RC (2014) A manual for Tilapia Business Management. water temperatures. International Journal of Agriculture and
Wallingford, Oxfordshire, UK, CABI. Biology, 11(3), 301–305.
Biswas AK, Takeuchi T (2003) Effects of photoperiod and feed- FAO (2018) Aquaculture Feed and Fertilizer Resources Informa-
ing interval on food intake and growth rate of Nile tilapia tion System, Available from URL: http://www.fao.org/fishery/
Oreochromis niloticus L. Fisheries Science 69(5): 1010–1016. affris/species-profiles/nile-tilapia/tables/en/ Accessed 11-7-
Brascamp EW, Merks JWM, Wilmink JBM (1985) Genotype 2018.
environment interaction in pig breeding programmes: meth- FAO (2016) FAO year Book 2014. Fishery and Aquaculture
ods of estimation and relevance of the estimates. Livestock Statistics.
Production Science 13(2): 135–146. Fessehaye Y, Kabir A, Bovenhuis H, Komen H (2006) Prediction
Brett JR, Groves TDD (1979) Physiological energetics. In: Hoar WS, of cannibalism in juvenile Oreochromis niloticus based on
Randall DJ, Brett JR (eds) In: Fish Physiology. Volume 8. Bioener- predator to prey weight ratio, and effects of age and stocking
getics and growth. pp. 280–344. Academic Press, London, UK. density. Aquaculture 255: 314–322.
Charo-Karisa H, Komen H, Reynolds S, Rezk MA, Ponzoni RW, Garcia F, Romera DM, Gozi KS, Onaka EM, Fonseca FS, Schalch
Bovenhuis H (2006) Genetic and environmental factors SHC et al. (2013) Stocking density of Nile tilapia in cages
affecting growth of Nile tilapia (Oreochromis niloticus) juve- placed in a hydroelectric reservoir. Aquaculture 410(411):
niles: modelling spatial correlations between hapas. Aquacul- 51–56.
ture 255(1): 586–596. Garc
ıa-Trejo JF, Pe~ na-Herrejon GA, Soto-Zaraz
ua GM, Mer-
Craig S (2009) Understanding Fish Nutrition, Feeds, and Feed- cado-Luna A, Alatorre-J
acome O, Rico-Garc
ıa E (2016) Effect
ing. In. Virginia Polytechnic Institute and State University, of stocking density on growth performance and oxygen con-
publication No. 420-256. Available from URL: https:// sumption of Nile tilapia (Oreochromis niloticus) under green-
vtechworks.lib.vt.edu/handle/10919/80712/ Accessed 11-7-2018 house conditions. Latin American Journal of Aquatic Research
Duan Y, Dong X, Zhang X, Miao Z (2011) Effects of dissolved 44(1): 177–183.
oxygen concentration and stocking density on the growth, Halver JE, Hardy RW (2002) Fish Nutrition. Academic Press, an
energy budget and body composition of juvenile Japanese imprint of Elsevier Science, San Diego, California.
flounder, Paralichthys olivaceus (Temminck et Schlegel). Huang Q, Huang K, Ma Y, Qin X, Wen Y, Sun L et al.
Aquaculture Research 42(3): 407–416. (2015) Feeding frequency and rate effects on growth and
Eknath AE, Bentsen HB, Ponzoni RW, Rye M, Nguyen NH, physiology of juvenile genetically improved farmed Nile
Thodesen J et al. (2007) Genetic improvement of farmed tila- tilapia. North American Journal of Aquaculture 77(4): 503–
pias: composition and genetic parameters of a synthetic base 512.
population of Oreochromis niloticus for selective breeding. Jobling M (1993) Bioenergetics: Feed intake and energy parti-
Aquaculture 273(1): 1–14. tioning, CHAPMAN & HALL.

Reviews in Aquaculture (2020) 12, 524–541

538 © 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd
 Determinants of yield gap in Nile tilapia farms

Jobling M (2003) The thermal growth coefficient (TGC) model Mohammad M, Uddin KB, Sanjib B, Yahia M, Muhammad Z,
of fish growth: a cautionary note. Aquaculture Research 34(7): Bai SC (2015) Effects of stocking density on growth, body
581–584. composition, yield and economic returns of monosex tilapia
Kabir KA, Schrama JW, Verreth JAJ, Phillips MJ, Verdegem (Oreochromis niloticus L.) under cage culture system in Kaptai
MCJ (2019) Effect of dietary protein to energy ratio on per- Lake of Bangladesh. Journal of Aquaculture Research and
formance of nile tilapia and food web enhancement in semi- Development, 6(8), 357–357.
intensive pond aquaculture. Aquaculture 499: 235–242. Mohammadi M, Sarsangi AH, Haghighi TD, Webster C, Rajabi-
Kamal AHMM, Mair GC (2005) Salinity tolerance in superior pour F, Mashaii N et al. (2014) Optimization of dietary pro-
genotypes of tilapia, Oreochromis niloticus, Oreochromis tein in all male Nile tilapia (Oreochromis niloticus) reared in
mossambicus and their hybrids. Aquaculture 247(1/4): 189–201. inland saline water. Animal Nutrition and Feed Technology 14
Kapinga I, Mlaponi E, Kasozi N (2014) Effect of stocking density (1): 91–99.
on the growth performance of sex reversed male Nile tilapia Mulder HA (2016) Genomic selection improves response to
(Oreochromis niloticus) under pond conditions in Tanzania. selection in resilience by exploiting genotype by environment
World Journal of Fish and Marine Sciences 6(2): 156–161. interactions. Frontiers in Genetics, 7(178): 1–11.
Kaya GK, Bilg€ uven M (2015) The effects of feeding frequency on Mulder HA, Bijma P (2005) Effects of genotype by environment
growth performance and proximate composition of young interaction on genetic gain in breeding programs. Journal of
Nile tilapia (Oreochromis niloticus L.). Ziraat Fak€
ultesi Dergisi, Animal Science 83(1): 49–61.
€
Uludag Universitesi, 29(1), 11–18. Mustapha MK, Oladokun OT, Salman MM, Adeniyi IA, Ojo D
Kim KW, Wang XJ, Bai SC (2002) Optimum dietary protein (2014) Does light duration (photoperiod) have an effect on the
level for maximum growth of juvenile olive flounder Par- mortality and welfare of cultured Oreochromis niloticus and Clar-
alichthys olivaceus (Temminck et Schlegel). Aquaculture ias gariepinus? Turkish Journal of Zoology 38(4): 466–470.
Research 33(9): 673–679. Neira R (2010) Breeding in Aquaculture Species: Genetic
Komen H, Trinh Quoc T (2014) Nile tilapia genetic improve- Improvement Programs in Developing Countries. 9th World
ment: achievements and future directions. In Proceedings of Congress on Genetics Applied to Livestock Production. Leip-
The 10th International symposium on Tilapia in Aquaculture zig, Germany.
(ISTA10) (2014), Jerusalem, Isreael, 6-10, October, 2014. pp. Ng WK, Romano N (2013) A review of the nutrition and feeding
1-9. management of farmed tilapia throughout the culture cycle.
Kpundeh MD, Qiang J, He J, Yang H, Xu P (2015) Effects of Reviews in Aquaculture 5(4): 220–254.
dietary protein levels on growth performance and haemato- Norman-L
opez A, Bjørndal T (2009) Is tilapia the same product
immunological parameters of juvenile genetically improved worldwide or are markets segmented? Aquaculture Economics
farmed tilapia (GIFT), Oreochromis niloticus. Aquaculture & Management 13(2): 138–154.
International 23(5): 1189–1201. NRC (2011) Nutrient Requirements of Fish and Shrimp. National
Lanna EAT, Bomfim MAD, Ribeiro FB, Quadros M (2016) Academies Press, Washington, D.C..
Feeding frequency of Nile tilapia fed rations supplemented Omasaki SK, Charo-Karisa H, Kahi AK, Komen H (2016a)
with amino acids. Revista Caatinga 29(2): 458–464. Genotype by environment interaction for harvest weight,
Lekang O-I (2013) Aquaculture Engineering. Wiley-Blackwell, growth rate and shape between monosex and mixed sex Nile
Chichester, West Sussex, UK. tilapia (Oreochromis niloticus). Aquaculture 458: 75–81.
Li D, Liu Z, Xie C (2012) Effect of stocking density on growth Omasaki SK, van Arendonk JAM, Kahi AK, Komen H (2016b)
and serum concentrations of thyroid hormones and cortisol Defining a breeding objective for Nile tilapia that takes into
in Amur sturgeon, Acipenser schrenckii. Fish Physiology and account the diversity of smallholder production systems. Jour-
Biochemistry 38(2): 511–520. nal of Animal Breeding and Genetics 133(5): 404–413.
Likongwe JS, Stecko TD, Stauffer JR Jr, Carline RF (1996) Com- Omasaki SK, Janssen K, Besson M, Komen H (2017) Economic
bined effects of water temperature and salinity on growth and values of growth rate, feed intake, feed conversion ratio, mor-
feed utilization of juvenile Nile tilapia Oreochromis niloticus tality and uniformity for Nile tilapia. Aquaculture 481:
(Linnaeus). Aquaculture 146(1–2): 37–46. 124–132.
Liu J, Li Z, Li X, Wang Y (2013) On-farm feed management Papoutsoglou SE, Karakatsouli N, Pizzonia G, Dalla C, Polissidis
practices for Nile tilapia (Oreochromis niloticus) in southern A, Papadopoulou-Paifoti Z (2006) Effects of rearing density
China. In Hasan MR and New MB (eds) On-farm feeding and on growth, brain neurotransmitters and liver fatty acid com-
feed management in aquaculture. FAO Fisheries and Aquacul- position of juvenile white sea bream Diplodus sargus L. Aqua-
ture Technical Paper No. 583. pp. 71–99. FAO, Rome. culture Research 37(1): 87–95.
Lovell T (1989) Nutrition and Feeding of Fish. MA, Springer, US, Ponzoni RW, Nguyen NH, Khaw HL, Hamzah A, Bakar KRA,
Boston. Yee HY (2011) Genetic improvement of Nile tilapia (Ore-
Mjoun K, Rosentrater K, Brown ML (2010) TILAPIA: Environ- ochromis niloticus) with special reference to the work con-
mental Biology and Nutritional Requirements. In Fact Sheets. ducted by the WorldFish Center with the GIFT strain. Reviews
Paper 164., http://openprairie.sdstate.edu/extension_fact/164. in Aquaculture 3(1): 27–41.

Reviews in Aquaculture (2020) 12, 524–541

© 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd 539
S. B. Mengistu et al.

Popma TJ, Lovshin LL (1996) Worldwide Prospects for Commer- genotypes that can thrive in a range of production systems?
cial Production of Tilapia. International Center for Aquacul- Scientific Reports 6: 21486.
ture and Aquatic Environments, Auburn, Ala.. Tran NTK, Dinh NT, Tin NH, Roem AJ, Schrama JW, Verreth
Popma T, Masser M (1999) Tilapia: Life history and biology. JAJ (2016) Interaction between dissolved oxygen concentra-
Southern Regional Aquaculture Center,SRAC publication No. tion and diet composition on growth, digestibility and intesti-
283, Stoneville, Ms, USA. nal health of Nile tilapia (Oreochromis niloticus). Aquaculture
R Core Team (2015) R: A language and environment for statistical 462: 101–108.
computing. R Foundation for statistical Computing, Vienna, Tran-Duy A, Schrama JW, van Dam AA, Verreth JAJ (2008)
Austria. Available from URL: http://www.R-project.org/ Effects of oxygen concentration and body weight on maxi-
Accessed 7-3-2018. mum feed intake, growth and hematological parameters of
Rana KJ, Hassan MR (2013) On-farm feeding and feed manage- Nile tilapia. Oreochromis niloticus. Aquaculture 275(1/4): 152–
ment practices for sustainable aquaculture production: an 162.
analysis of case studies from selected Asian and African coun- Tran-Duy A, van Dam AA, Schrama JW (2012) Feed intake,
tries. On-farm feeding and feed management in aquaculture. growth and metabolism of Nile tilapia (Oreochromis niloticus)
FAO Fisheries and Aquaculture Technical Paper No. 583 No. in relation to dissolved oxygen concentration. Aquaculture
583. pp. 21-67. Rome, FAO. Research 43(5): 730–744.
Randall DJ, Tsui TK (2002) Ammonia toxicity in fish. Marine Trọng TQ, Mulder HA, van Arendonk JAM, Komen H (2013)
pollution bulletin 45(1–12): 17–23. Heritability and genotype by environment interaction esti-
Ridha MT (2006) Comparative study of growth performance of mates for harvest weight, growth rate, and shape of Nile tila-
three strains of Nile tilapia, Oreochromis niloticus, L. at two pia (Oreochromis niloticus) grown in river cage and VAC in
stocking densities. Aquaculture Research, 37(2), 172–179. Vietnam. Aquaculture 384–387: 119–127.
Robertson A (1959) The sampling variance of the genetic corre- de Verdal H, Benzie JA, Komen H, Quillet E, Vandeputte M,
lation coefficient. Biometrics 15(3): 469–485. Chatain B et al. (2018) Improving feed efficiency in fish using
Romana-Eguiaa MRR, Larona MA, Catacutan MR (2013) On- selective breeding: a review. Reviews in Aquaculture. 10(4):
farm feed management practices for Nile tilapia (Oreochromis 833–851.
niloticus) in the Philippines. In M.R. Hasan and M.B. New, Wang T, Lefevre S, Thanh Huong DT, Cong NV, Bayley M,
eds. On-farm feeding and feed management in aquaculture. Richards JG et al. (2009) Chapter 8 the effects of hypoxia on
FAO Fisheries and Aquaculture Technical Paper No. 583. pp. growth and digestion. In: Fish Physiology. pp. 361–396. Aca-
131-158. Rome, FAO. demic Press, London, UK.
Sae-Lim P, Gjerde B, Nielsen HM, Mulder H, Kause A (2016) A Watanabe WO, Kuo CM, Huang MC (1985). The ontogeny of
review of genotype-by-environment interaction and micro- salinity tolerance in the tilapias Oreochromis aureus, O.
environmental sensitivity in aquaculture species. Reviews in niloticus, and an O. mossambicus 9 O. niloticus hybrid,
Aquaculture 8(4): 369–393. spawned and reared in freshwater. Aquaculture, 47(4), 353–
Santiago CB, Aldaba MB, Reyes OS (1987) Influence of feeding 367.
rate and diet form on growth and survival of Nile tilapia (Ore- Yakubu AF, Okonji VA, Nwogu NA, Olaji ED, Ajiboye OO,
ochromis niloticus) fry. Aquaculture 64(4): 277–282. Adams TE (2013) Effect of stocking density on survival
Stickney RR (2017) Aquaculture: An Introductory Text. CABI, and body composition of Nile tilapia (Oreochromis niloti-
Wallingford, Oxfordshire, UK. cus) fed MULTI feed and NIOMR feed in semi flow-
Sweilum MA, Abdella MM, El-Din SAS (2005) Effect of dietary through culture system. Journal of Natural Sciences Research
protein-energy levels and fish initial sizes on growth rate, 3(14): 29–37.
development and production of Nile tilapia. Oreochromis Yi Y, Lin CK, Diana JS (1996) Influence of Nile tilapia (Ore-
niloticus L. Aquaculture Research 36(14): 1414–1421. ochromis niloticus) stocking density in cages on their growth
Thoa NP, Ninh NH, Knibb W, Nguyen NH (2016) Does selec- and yield in cages and in ponds containing the cages. Aquacul-
tion in a challenging environment produce Nile tilapia ture 146(3/4): 205–215.

Reviews in Aquaculture (2020) 12, 524–541

540 © 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd
 Determinants of yield gap in Nile tilapia farms

APPENDIX 1: List of studies that were the source of data included for each analysis

Model 1.
Investigated factor: FE; number of studies/number of data records: 28/179

Abdel-Tawwab et al. (2014)1, Abdel-Tawwab et al. (2015)1, Abou et al. (2007)1, Al-Hafedh (1999)1, Alhassan et al. (2012)1, 2, Ali et al. (2008)1,
Azaza et al. (2008)1, Azaza et al. (2013)1, Azaza et al. (2015)1, Azevedo et al. (2015), Bahnasawy (2009), 2, Biswas and Takeuchi (2003)1, 2, El-
Sayed and Teshima (1992), El-Sherif and El-Feky (2009)1,2, Garcia et al. (2013)1, Huang et al. (2015) 1, 2, Kamal and Mair (2005)1, Kapinga et al.
(2014)1, 2, Kaya and Bilgu€ven (2015)1, 2, Kpundeh et al. (2015)2, Lanna et al. (2016)1, Mohammad et al. (2015)1, Mohammadi et al. (2014)1, Ridha
(2006)2, Santiago et al. (1987)1, 2, Sweilum et al. (2005)1, 2, Tran-Duy et al. (2008)1, Yi et al. (1996)1
Model 2.
Investigated factor: FE; number of studies/number of data records: 23/141
Studies mentioned for model 1 above that are marked with superscript 1
Model 3
Investigated factor: FE; number of studies/number of data records: 11/67
Studies mentioned for model 1 above that are marked with superscript 2
Model 4
Investigated factor: Survival; number of studies/number of data records: 29/187
Abdel-Tawwab et al. (2014, 2015), Abou et al. (2007), Al-Hafedh (1999), Alhassan et al. (2012), Azaza et al. (2008, 2015), Azevedo et al. (2015),
Bahnasawy (2009), Biswas and Takeuchi (2003) El-Sayed and Teshima (1992), El-Sherif and El-Feky (2009), Garcia et al. (2013), Garc
ıa-Trejo et al.
(2016), Huang et al. (2015), Kamal and Mair (2005), Kapinga et al. (2014), Kaya and Bilgu €ven (2015), Kpundeh et al. (2015), Lanna et al. (2016),
Likongwe et al. (1996), Mohammad et al. (2015), Mustapha et al. (2014), Ridha (2006), Santiago et al. (1987), Sweilum et al. (2005), Tran-Duy
et al. (2008), Yakubu et al. (2013), Yi et al. (1996)
Model 5
Investigated factor: TGC; number of studies/number of data records: 29/192
Abdel-Tawwab et al. (2014)3, Abdel-Tawwab et al. (2015)3, Abou et al. (2007)3, Al-Hafedh (1999)3, Alhassan et al. (2012)3, 4, Ali et al. (2008)3,
Azaza et al. (2008)3, Azaza et al. (2013)3, Azaza et al. (2015), Bahnasawy (2009)3, 4 Biswas and Takeuchi (2003)3, 4 El-Sayed and Teshima (1992),
El-Sherif and El-Feky (2009)3, 4, Garcia et al. (2013)3,4 Garc
ıa-Trejo et al. (2016), Huang et al. (2015)3, 4, Kapinga et al. (2014)3, 4, Kaya and
€ven (2015)3, 4, Kpundeh et al. (2015)4, Lanna et al. (2016)3, Likongwe et al. (1996)3, 4, Mohammad et al. (2015)3, Mohammadi et al. (2014)3
Bilgu
Ridha (2006)4, Santiago et al. (1987)3,4, Sweilum et al. (2005)3,4, Tran-Duy et al. (2008)3, Yakubu et al. (2013)3,4, Yi et al. (1996)3
Model 6
Investigated factor: TGC; number of studies/number of data records: 24/155
Studies mentioned for model 6 above that are marked with superscript 3
Model 7
Investigated factor: TGC; number of studies/number of data records: 14/86
Studies mentioned for model 6 above that are marked with superscript 4

Reviews in Aquaculture (2020) 12, 524–541

© 2019 The Authors. Reviews in Aquaculture Published by John Wiley & Sons Australia, Ltd 541