Nutrition for juvenile African cichlids

MISSING IMAGE

Material Information

Title:
Nutrition for juvenile African cichlids the effects of varying dietary protein and energy levels on growth performance and liver condition
Physical Description:
x, 134 leaves : ill. ; 29 cm.
Language:
English
Creator:
Royes, Juli-Anne B
Publication Date:

Subjects

Subjects / Keywords:
Fisheries and Aquatic Sciences thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Fisheries and Aquatic Sciences -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2004.
Bibliography:
Includes bibliographical references.
Statement of Responsibility:
by Juli-Anne B. Royes.
General Note:
Printout.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 003100972
oclc - 706798739
System ID:
AA00003586:00001

Full Text

















NUTRITION FOR JUVENILE AFRICAN CICHLIDS: THE EFFECTS OF VARYING
DIETARY PROTEIN AND ENERGY LEVELS ON GROWTH PERFORMANCE AND
LIVER CONDITION












By

JULI-ANNE B. ROYES


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2004































This is dedicated to my parents, Carol and Wesley; to my sister Tricia; and to my
husband Riccardo. Thank you for your love and encouragement.














ACKNOWLEDGMENTS

I thank the members of my supervisory committee (Ruth Francis Floyd. Chair:

Debra Murie, Cochair; and members Richard Miles, Roy Yanong. Scott Terrell and

Charles Adams) for their guidance and support throughout my research. Sincere

gratitude and appreciation are extended to Drs. Debra Murie and D. Allen Davis (Auburn

University, Department of Fisheries & Allied Aquaculture) for their many contributions

in the preparation of this manuscript; and to Drs. Richard Miles and Frank Chapman who

were always available for consultation, guidance and words of encouragement throughout

the course of this program.

I also thank African cichlid farmers R. Biro. P. Radice. L. DeMason. S. Cartwright.

and W. Warner (and other members of the Florida Tropical Fish Farms Association) for

their help. Special thanks go to Mrs. R. E. Klinger and to Misters J. Holloway. R. Russo,

R. Fethiere, F. Robbins, P. Hamilton, V. Sampath, C. Martinez and S. Graves for their

assistance with fieldwork and conducting laboratory analyses. Special thanks are

extended to fellow graduate students Jeff Hill, Joel Carlin. Alejandro Ruiz. and Jesus

Venero for their friendship and understanding and invaluable assistance throughout the

different phases of this study. Lastly to Dr. A. Ahn and to Mr. R. Reid (Hartz Mountain

Corp. & LM Animal Farms) for enabling me to put this work into practice.

This project was supported by a grant from USDA-CSREES (JBR, RFF. DJM) and

a University of Florida (UF) doctoral fellowship administered by the University of

Florida [UF], Institute of Food and Agricultural Science.














TABLE OF CONTENTS

Page

ACKNOWLEDGMENTS .......................................................... iii

L IST O F T A B LE S ...................................................... ................................................ vi

L IST O F FIG U R E S ................................................................................................... viii

A B ST R A C T ................. .... ................................................................ ....... .............. ix

CHAPTER

1 INTRODUCTION ...................................................................... .................

Tropical Ornamental Fish Farming Industry ........................................................ 1
African Cichlid Industry in Florida........................... .............. ....... 2
Feeding of A frican C ichlids .......................................................... .................... 4
N nutrition for A frican Cichlids.................... ............................... .......................... 11

2 EFFECTS OF TWO COMMERCIALLY PREPARED FEEDS ON GROWTH
PERFORMANCE AND LIVER COMPOSITION OF JUVENILE AFRICAN
CICHLIDS (Pseudotropheus socolofi AND Haplochromis ahli)

Introduction..... ......................... ..................... ............. ... 15
M materials and M ethods........... ....................... ........ ..................................... 17
Results...... ............... ........ ..... .................... 20
D discussion ........... .. ............ ........ ....... ... ........................... .... ....... ........... 2 1

3 EFFECTS OF DIETARY PROTEIN LEVELS ON GROWTH
PERFORMANCE, PROTEIN EFFICIENCY AND IEPATOCYTE CHANGES
IN JUVENILE AFRICAN CICHLIDS (Pseidolropheus socolofi)

Introduction..................... ........ ................ .... .. .......32
M materials and M ethods.........................................................................................34
Results .......... .... .... ... ........ ..... ... ................... ...38
D discussion .................................................. .. ...... ........ ..... 40









4 EFFECTS OF VARYING DIETARY PROTEIN AND ENERGY
LEVELS ON GROWTH PERFORMANCE AND HEPATOCYTE
CHANGES IN JUVENILE AFRICAN CICHLIDS (Pseudotropheus
socolofi AND Haplochromis ahli)

Introduction .................... .................... ..... ........... ................................. 49
M materials and M ethods.......................................... ........................................50
Results.. ............. ................ .... .. .............................. ..... 55
D discussion ............ .. ............ ... .. ......... .... ............. ... .................. ........... ..... 60

5 FORMULATION OF A MODEL DIET FOR THE COMMERCIAL
PRODUCTION OF JUVENILE AFRICAN CICHLID FISHES

Introduction ........................................................................... .. ...................... .... 77
Nutrient Requirements and Feed Ingredients ...................................................78
Feed Formulation and Preparation of a Juvenile African Cichlid Diet .................85
Considerations for Preparation of Farm-Made Feeds versus the Use of
Com m ercial Feeds by Farm ers ...................... ............................................... 89
The Potential for the Commercial Manufacturing of African Cichlid Diets ........92

6 C O N C L U S IO N ......................................................................... ........................ 106

APPENDIX

A SURVEY OF THE NUTRITIONAL MANAGEMENT OF AFRICAN
CICHLID FARMS IN SOUTH FLORIDA SPRING 2000..............................113

B SU R V E Y FO R M ......................................................... ............................... ... 117

C DIET ANALYSES........................................................................... .... 121

LIST O F REFEREN C ES ..................................................... .............................. 123

BIOGRAPHICAL SKETCH..................................................... .................... 134














LIST OF TABLES


Tables Page

2-1. Proximate composition of two commercial diets used in feeding
trials for juvenile African cichlids ............ ..................................... .................28

2-2. Mean growth performance and hepatocyte changes of juvenile
African cichlids fed two commercial diets for 12 weeks......................................29

3-1. Formulation and composition (g /100 g dry weight ) of
experim ental diets ..... ..................................................................................43

3-2. Growth parameters and hepatocyte changes of juvenile P. socolofi
after 10 weeks on experimental diets. ................. ........... .................... 44

3-3. Carcass composition of P. socolofi after 10 weeks of feeding diets
with increasing protein content................................ .......... ................. 45

4-1. Formulation and composition (g / 100 g dry weight) of
experimental diets. ..................................................................................... 67

4-2. Growth parameters of juvenile P. socolofi after 8 weeks on
experim ental diets ............... ...................................................... ..........68

4-3. Growth parameters of juvenile H. ahli after 8 weeks on
experimental diets. ...................................................................... 69

4-4. Whole body composition (%) and protein conversion efficiency
(%) of P. socolofi and H. ahli after 8 weeks on experimental diets.........................70

4-5. Hepatosomatic index and liver lipid composition of P. socolofi and
H. ahli after 8 weeks on experim ental diets. ......................................................... 71

5-1. General amounts of nutrients incorporated into diets for juvenile
A frican cichlids ...................... ................................ .............. ............................96

5-2. Nutrient composition and costs per pound of feed ingredients for
the formulation a commercial feed for juvenile African cichlids........................97

5-3. An Excel spreadsheet for the formulation of a model diet for
juvenile African cichlids ........................ ............................................98









5-4. The output of a model diet for juvenile African cichlids .......................................99

5-5. Specifications sheet for the commercial manufacture of a slowly sinking feed for
juvenile African cichlids ..................................... ..... ...... ................. 100

5-6. A comparison of the nutrient analyses and costs between the recommended cichlid
diet and a trout diet being used presently ......................................... ................. 101

5-7. The costs per pound of raw ingredients available............................................... 102

5-8. The output of a model diet for juvenile African cichlids to be made on the farm .103

5-9. Options for equipment used to make farm-made feeds .....................................104

C-1. Nutrient analyses of a model commercial feed for juvenile African cichlids........122














LIST OF FIGURES


Figures Page

2-1. Mean weight of juvenile African cichlids (P. socolofi and H. ahli) fed
commercial Diet 1 (high-energy; high-lipid pellet) or Diet 2 (low-energy:
low -lipid flake)......................................................... ......................... .... 30

2-2. Histological liver sections from P. socolofi and H. ahli fed
experim ental diets for 12 w eeks ........................................................................... 1

3-1. Mean growth response of P. socolofi fed experimental diets with
different dietary crude protein levels (%) for 10 weeks. .....................................46

3-2 Mean final weight of P. socolofi fed diets with varying dietary
crude protein levels (%) for 10 weeks. ........................................ ..............47

3-2. Protein conversion efficiency (PCE) and Protein efficiency ratio (PER) of P.
socolofi fed experimental diets for 10 weeks........................................48

4-1. Mean growth of P. socolofi fed experimental diets with varying dietary protein
(% ) and lipid levels (% ) for 8 weeks ........................................ .................72

4-2. Mean growth of H. ahli fed experimental diets with different dietary protein
levels (% ) for 8 w eeks............................................. .................. ..............73

4-3. Histological liver sections from P. socolofi fed experimental diets for 8 weeks...74

4-4. Histological liver sections from I. ahli fed experimental diets for
8 weeks .................... ....... .......... ...... .......... ............75

4-5. Periodic acid Schiff stained liver sections of P. socolofi and H. ahli fed
experim ental diets for 8 w eeks ................................................ ..................76

5-1. Methods of distribution of manufactured feeds being used...............................105








Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

NUTRITION FOR JUVENILE AFRICAN CICHLIDS: THE EFFECTS OF VARYING
DIETARY PROTEIN AND ENERGY LEVELS ON GROWTH PERFORMANCE AND
LIVER CONDITION

By

Juli-Anne B. Royes

August 2004

Chair: Ruth Francis Floyd
Cochair: Debra J. Murie
Major Department: Fisheries and Aquatic Sciences

The objectives of this dissertation were (1) to determine the minimum requirements

for protein and energy for juvenile African cichlids needed to provide normal growth

without pathological changes in liver cells; and (2) to formulate a cost-effective model

diet. A 12-week feeding trial was conducted to evaluate growth performance and

hepatocyte changes of juvenile African cichlids (Pseudotropheusl socolof and

Haplochromis ahli) fed commercial diets used on cichlid farms in South Florida. Growth

was significantly greater (P < 0.05) for both species fed a trout starter pellet (52% crude

protein, 17% lipid) than for fish on a mixed flake feed (45% crude protein, 7% lipid).

This trial showed that the trout starter pellet may be suitable for commercial production

of juvenile African cichlids up to 12 weeks of age; but prolonged feeding may result in

excess lipid deposition and necrosis of the liver.

A 10-week trial was conducted to observe the effects of varying protein levels on

growth performance and changes in hepatocytes of juvenile P. socolofi (0.09 g: 1.8 cm).

Six diets were formulated with protein levels ranging from 32.5 to 58.8%. Survival was

greater than 95% for all treatments. The minimum range of dietary protein to produce








optimum growth of P. socolofi without significant changes in liver lipid levels or

pathology was determined to be 36.9 to 42.5%.

A third feeding trial was conducted for 8 weeks to observe the effects of varying

energy (ranging from 4,000 to 5,300 kcal/kg) at two protein levels (36 and 50%) on

growth performance and hepatocyte change of P. socolofi (0.89 g; 4.1 cm) and H. ahli

(2.04 g; 5.65 cm). Based on growth performance and liver condition, a diet containing

36% protein, 9.5% lipid at an energy level of 4000 kcal/kg, was determined to be suitable

for commercial production of these juvenile African cichlids. With these requirements.

the formulation of a practical diet became possible. This diet should improve nutrient

use. provide growth rates equivalent to or better than growth rates of commercial feeds

currently in use decrease the level of liver and carcass fat deposition. and maintain good

water quality.














CHAPTER 1
INTRODUCTION

Tropical Ornamental Fish Farming Industry

Tropical aquarium fish are cultured primarily for the home aquarium. with many

species and varieties being traded from both freshwater and marine origin. According to

Chapman (2000), approximately 95% offish traded are from farm-raised freshwater

species (with over 1,000 species in about 100 families represented in the ornamental fish

trade at any one time). Considered a luxury item, these fish are kept as a hobby and are

popular pets. Worldwide, the aquarium industry is worth US $1.5 billion, of which the

United States has the largest share (Chapman 2000). In 2000. sales of ornamental fish in

the U.S. were estimated to be more than $350 million annually (Martin 2000).

In North America, most fresh-water ornamental fish are predominantly produced

in Florida. Tropical fish farming in Florida has been a viable venture since the 1930s.

with the industry starting in Tampa, Florida (Socolof 1996). However. only within the

last 30 years has major expansion occurred, with farms now producing over 700 varieties

(Watson and Shireman 1996). It is estimated that Florida farms produce approximately

80% of the domestically bred pet fish. Eighty percent of this production is in the counties

surrounding Tampa (including the counties of Hillsoborough. Polk, Manatee and

Sarasota). Other areas that produce pet fish include the east coast of Florida (from Vero

Beach to Miami). These counties have convenient access to major airports in Tampa.

Miami and Orlando for shipping. Today, tropical pet-fish farming continues to be the

leading segment (in sales volume) in Florida's aquaculture industry (Florida Agricultural









Statistics Service [FASS] 2002). In 2001, net sales of tropical fish totaled $42.4 million

from 160 growers in Florida (representing 43% of total aquaculture sales in Florida)

(FASS 2002). The value of exports from Florida in 2000 was $4.2 million with Asian

countries (in particular Hong Kong) representing the largest market. The availability of

adequate ground water, together with the presence of a mild tropical climate all year

round, gives Florida a competitive advantage over other geographic regions in the

continental United States. Ornamental fish comprise two broad categories based on their

reproductive strategies: live bearers and egg layers. Of the total net sales in 2001 (142.4

million), live bearers (e.g. swordtails, guppies) totaled $13.2 million while egg layers

(e.g. cichlids, barbs) amounted to $29.3 million (FASS 2002).

African Cichlid Industry in Florida

African cichlids are one of the most popular groups of ornamental fishes sold.

Their diversity, coloration, behavioral characteristics and size have appealed to

generations of aquarists (Goldstein 1987). Many African cichlids are brightly colored

with various shades of blue, yellow, orange and red. Some have various contrasting

colored spots, dark bars and stripes or iridescent color schemes. The history of African

cichlid classification and biology has been documented by several biologists and

hobbyists (Fryer 1959; Fyer and les 1972; Keenlyside 1991; Axelrod and Burgess 1993:

DeMason 1995). It was not until the 1950s, however, that these fish became of interest to

the aquarium industry; and by the early 1980s they were intensively produced on farms in

Florida (personal communication, Rick Biro, African cichlid farmer). Most of the

African cichlids known in the ornamental fish trade are collected and exported from the

coastline of Lake Malawi in the country of Malawi. In the 1990s, the coastline of Iake








Malawi on the Tanzanian border was explored (DeMason 1995); and with the consequent

exploration of Lake Tanganyika in Tanzania, the vast array of fishes available to the

industry was increased. An undocumented smaller number of cichlids in the market also

originate from Lake Victoria in Uganda. It has been suggested that there may be 500-

1000 African cichlid species present in Lake Malawi with 99% being endemic (Ribbinik

1991; Stiassay 1991) and over 200 species in Lake Tanganyika (DeMason 1995). The

number of species and hybrids traded in the market are too numerous to count but within

these groups it is suggested that over 500 species from these African Lakes have been

captured and bred (Klinger et al. 2001).

Many of the rock-associated fishes known by the Malawi as 'mbuna'. species

which include the genera Labeotropheus, Labidochromnis, Melanochromis and

Pseudotropheus, are popular species grown in Florida (Appendix A). Species found in

open waters that are active predators are represented by the genera Copadichro&nis,

Haplochromis, Nimbochromis and Sciaenochromis. Those that may be found over sand

bottoms as well as the sand-rock surface in Lake Malawi include the genera Aulonocara,

Oreochromis and Tilapia (of which many species are cultivated for human consumption)

(Axelrod and Burgess 1993).

Today it is estimated that there are about 16 farms that are the primary producers

of African cichlids (Appendix A) located mainly in Dade and Hillsborough counties of

Florida. Water quality parameters in these areas are similar to those found in Lakes

Malawi and Tanganyika: high temperatures (26 to 300C), alkalinity (50 to 80 ppm).

hardness (180 to 300 ppm) and pH (7.5 to 8.5) making it possible to grow these fishes in

Florida. Farms within these counties (and others in South Florida) have a mixed culture








of fishes, which include African cichlids (Appendix A). Most farms are family-owned

and family-operated and combine the use of indoor and outdoor facilities for fish

production. Traditionally, fish are produced in outdoor concrete ponds fed by

continuous-water flow artesian wells. In recent years, farmers have increased production

of African cichlids in indoor facilities. As a result of their small size and high individual

value, these fish can be raised intensively and profitably indoors. Juveniles in production

are maintained primarily in concrete vats (e.g., burial vaults) and glass aquaria of various

sizes. The use of plastic and fiberglass tanks is also common. In general, it takes about 2

to 6 months to attain a market-ready fish of 4 to 5 cm (1.5 to 2 inches) for most species;

and 7.5 cm (3 inches) for Haplochromine spp. Grow-out to adult size takes about 5 to 8

months (personal communication, African cichlid farmers). Success of fish production is

related to the quality of food fed and proper farm management practices. Management

practices in these intensive systems are designed to maintain good quality water and to

provide proper nutrition (Chapman 2000). Only under these conditions can these fish be

raised profitably at high densities, be in good health, and attain proper coloration. These

fish command among the highest prices of aquacultural products, often several hundred

dollars per kilogram. Wholesale prices can range from $0.75 to $100.00 (depending on

size and species) for an individual fish. From a survey taken in 2002, the annual fish

production from Dade county alone was estimated to be $3M (Appendix A) and the total

annual sales from production of African cichlids from all the farms in Florida was $10M.

Feeding of African Cichlids

Little information exists on the nutritional requirements of African cichlids for the

pet industry. Knowledge of their dietary requirements is largely anecdotal and has









evolved primarily from trial and error (by individual farmers and feed manufacturing

companies). African cichlids are mouth brooders and produce yolk-sac bearing larvae

that are able to commence eating commercial feeds about 3-6 days after hatching

(African cichlid farmers, personal communication). After absorption of the yolk sac.

juveniles readily accept commercially formulated feeds (200 to 400 p. pelleted feeds) and

are supplemented with flakes (Spirulina and mixed) and live foods. including brine

shrimp (Artemia salina). The ability to eat artificial feeds at the initiation of exogenous

feeding has been a major factor in the success of culturing African cichlids. Both farm-

made feeds and feed mixtures from commercial sources are used for feeding fish cultured

indoors. Commercial feeds purchased are usually feeds developed for the salmon and

trout industries and contain 45 to 55% crude protein (CP) and 12 to 17% fat. Farm-made

feeds usually consist of high-protein diets (45 to 60% CP). Proteins from animal sources

(such as fish meal, shrimp meal, chicken eggs and liver) may be used on the farm

(Appendix A). Feed is delivered primarily by hand; and is fed at least twice a day.

With the intensification in production and the increased use of recirculating and

flow-through systems, healthy, fast-growing fish is the primary interest of farmers in

order to increase the turnover rate, thereby boosting profits. Artificial diets should

therefore be well-balanced to contain all the essential nutrients, energy and pigments to

achieve this goal. Although numerous diets are available for purchase by the individual

pet-fish owner, presently no feeds are available for the specific commercial production of

African cichlids. Thus fish of all sizes are fed the same commercial diets regardless of

their natural feeding ecology and stage in the life cycle. Natural feeding habits of








cichlids are diverse; and specializations range from strict herbivores to strict piscivores

(Keenlyside 1991, Axelrod and Burgess 1993, DeMason 1995).

Evidence of Nutritional Diseases in African Cichlids

Under production conditions, it is difficult to define nutritional diseases in

absolute terms. Nutritionally incomplete diets will render fish more susceptible to

infections that are more clinically obvious and make the underlying nutritional basis for

the diseased condition difficult to define (Roberts 2002). In the late 1990s, suspicion of

nutritional imbalances were recorded for fish that were sent to the University of Florida's

Fish Diagnostic Laboratory (IFFDL). Affected fish appeared to have unrelated

infectious diseases and were observed to have extensive amounts of visceral fat and very

large fatty livers (Ruth Francis Floyd, personal communication). In that same period.

major producers of ornamental African cichlids suffered a significant increase in fish

mortality. Upon internal examination, livers were tan in color, fatty with hemmorraghic

streaks, and necrotic. Histologic examination showed heavy vacuolation, cloudy

swelling of cells, and necrosis of the hepatocytes in the liver and hepatopancreas. Ceroid

deposits were observed in the liver, pancreas, and spleen. Although the feeds of the fish

submitted to the lab were not analyzed, it was suspected that the diets used \were high in

fat and contributed to the disease signs. For example, a diet containing excess fat can

cause fatty liver syndrome (Ferguson 1989; Roberts 2002). Also. diets with more than

10% fat can undergo rancidity (Gatlin III 2002) especially in a high-temperature climate.

(as that found in south Florida). Feeding rancid feeds may cause an abundance of ceroid-

laden macrophages in the visceral organs (Ferguson 1989; Roberts 2002). Farmers were

encouraged to change diets, feed a diet lower in fat (< 10%), and supplement with a








vitamin premix; after which the signs decreased.

Other evidence of nutritional disease of African cichlids was observed by Noga

(1986) and Klinger et al. (2001). Klinger et al. (2001) examined Lake Victorian cichlid

(Haplochromis (Prognathrochromis) perrieri) females with waxy enlarged ovaries,

containing intermittent large fluid-filled sacs; and males with internal multifocal

granulomas. Both sexes had ceroid deposits in all organs; and in females, the normal

ovarian tissue had been replaced with ceroid deposits and was necrotic. It was suspected

that ceroidosis in all organs observed resulted from lipid oxidation due to feed rancidity.

The increase in lipid production in the ovaries could also have created enlargement and

dysfunction of the tissue. Multifocal granulomas in tissues found in the Lake Victoria

cichlids (Klinger et al. 2001), also suggested a nutritional association first proposed by

Noga (1986) where he found diet-associated granulomas in response to African cichlids

fed a tropical fish flake food.

Malawi Bloat, (a common term used to describe a poorly understood condition)

has been observed in captive African cichlids (Ferguson et al. 1985; Dixon et al. 1997;

Smith 2000). It has been suggested that fish specifically native to Lake Malawi (although

not confined only to members of this group) are prone to the condition. Affected fish

develop abdominal distention, become lethargic, and may stop eating or may visibly spit

food out of their mouths and produce long fecal casts (Dixon et al. 1997). Over the years,

Malawi bloat has been ascribed to a number of causes, including bacterial and viral

infections, and metabolic and nutritional disorders (Andrews et al. 1988). The condition

does not respond to treatment with generally available antibiotics; and has therefore been

described primarily as a nutritional problem (Dixon et al. 1997; Smith 2000). Dixon et









al. (1997) suggested that an imbalance in the diet (such as excess starch) might disrupt

the normal bacterial flora in the gut, causing an increase in pathogenic bacteria such as

Clostridium difficile.

Digestive Physiology

Categorization of a cichlid species according to the type of food ingested in nature

suggests that there are differences in the structural and physiological characteristics of the

digestive system among species, and in each of its major life stages. For example, the

structure of the stomach and the intestinal tract varies according to the type of food the

fish eats. In herbivorous species, the length of the alimentary canal is several times the

length of the body (Horn 1997). This allows enough time for the digestion of plant tissue

which is more difficult than animal tissue to digest. Herbivores must feed continuously

and consume more food than carnivores in order to meet their energy needs. The natural

diets of these fish are rich in plant proteins and lipids that provide the essential nutrients

and carotenoids. In contrast, the alimentary canals of piscivorous fish are short with very

well-developed stomachs. These fish have a preference for animal proteins; and require a

higher lipid content for energy, as they have difficulty digesting high concentrations of

carbohydrates (Horn 1997, Rust 2002).

Feeding ecology. Generally, the diet of the herbivorous fish consists of 95'.,

plant-based material that could range from microscopic filamentous algae or diatoms to

the leaves of large plants. For example, members of the genus Tilapia feed on blue-green

algae and diatoms. It was found that T nilotica consumes two types of blue-green algae:

Spirulina and Anabaenopsis (Fryer and Iles 1972). Carnivores may be categorized into

several groups: piscivores (95% of the diet comprises other fishes): insectivores; and









more specialized feeders (mollusk-eaters, scale-eaters) and planktivores (IHaplochromis

and the young of many other species) (Fryer and Iles 1972).

Most cichlids are actively omnivorous and therefore consume both plant and

animal material. In many cases, feeding categories are not very discrete and the overlap

among diets is considerable. Depending on availability, some omnivorous species of

cichlids may tend to be more herbivorous and consume more plant material while others

are carnivorous and consume a higher percentage of other fishes, insects or other

invertebrates. This diversity of dietary feeding preferences makes it difficult to find a

single feed for the large numbers of cichlids species that are commercially cultured.

Hence, many farmers feed all species the same diet. However, it was speculated that fish

performance on these diets depends on whether the species was primarily herbivorous.

omnivorous, or carnivorous in its natural habitat.

To determine if there is a diet, species-specific effect on growth performance and

general health, experiments in Chapters 2, 3, and 4, were performed using an

herbivorous/omnivorous species (Pseudotropheus socolofi) and a carnivorous species

(Haplochromis ahli). These species were selected based on their popularity, as suggested

by the survey (Appendix A) to be used as representative models of an omnivorous and a

carnivorous fish, respectively. Originally from Point Mozambique in Malawi. P. soco/ofi

is powder blue in color and can attain a maximum length of 7.5 cm (3 inches) when adult

(Axelrod and Burgess 1993; Smith 2000). They preferentially feed on rock algae.

Haplochromis ahli, commonly known as the electric blue hap, is an example of a

piscivorous fish. It is usually larger than P. socolofi and can attain an adult size of

approximately 15 cm (6 to 7 inches) (Smith 2000).









Structure and function of the teleost liver. The fish liver, as the liver of other

vertebrates, plays a key role in the metabolism of proteins, lipids and carbohydrates.

Ventrally located in the cranial region of the coelomic cavity, the liver is generally

reddish-brown, tending toward yellow when fat storage is high (Ferguson 1989: Roberts

2002; Rust 2002). Liver hepatocytes make up most of the liver volume and can contain

variable amounts of glycogen and lipid. Changes in the gross and histological

morphology (as well as the amount of glycogen and lipid in the liver) are often related to

health, diet, toxin and energy status of the fish (Ferguson 1989; Rust 2002).

The liver serves as the primary storage organ for lipid in African cichlids. It has

been observed that the liver of ornamental fish produced in captivity and fed artificial

feeds tend to have a pale color due to lipid accumulation (Ferguson 1989; Roberts 2002).

If the amount of stored lipid is excessive, the liver may become enlarged, and have a

rounded appearance, rather than sharp edges. The mean hepatosomatic index (HSI) value

is used as an index to determine the levels of fat deposition in the liver. It is defined as

the percent of wet liver weight divided by the wet weight of the whole fish. It is species-

specific and correlates with the quantity and quality of food eaten (Brusle and Anadon

1996). The presence of excess fat in the liver lowers the specific gravity of the tissue and

the liver therefore floats when immersed in formalin. Small droplets of fat from the liver

surface may also be seen floating on the surface of the fixative. An extensive

accumulation of lipid makes the liver friable, that is, reduced in strength and toughness.

Tissues may fall apart to the touch as the tissues offer very little resilience (King and

Alroy 1997).









The nucleus of normal fish hepatocytes is e-'.. I illy round to ovoid and centrally

located. The cytoplasm of the hepatocytes may include normal stores of neutral fat

droplets that appear as empty, clear, round spaces and are widely distributed throughout

the liver tissue. With excess stores, fat droplets are concentrated at the center of the cell

and shift the nucleus and organelles to the periphery. Accumulation of neutral fat within

cells also causes the cells to expand. Cells can become so distended that they rupture.

spilling fat droplets into the surrounding tissue (King and Alroy 1997). Although

pathology of the liver due to metabolic diseases may be reversible, a prolonged diet high

in fat or starch can lead to decreased immunity and development of opportunistic

infections (King and Alroy 1997). Fatty change is generally regarded as a reversible

form of cell injury, since removal of the cause results in the fatty liver returning to the

normal state. If the underlying cause is not corrected, affected cells become irreversibly

damaged and die (King and Alroy 1997). Hepatic alterations that can occur quickly due

to feeding an unbalanced diet can therefore be used as an indication of changes in the

health status of the fish.

Nutrition for African Cichlids

Due to a lack of scientific information on the nutrition of ornamental African

cichlids, individual farmers have developed their own feeding practices for Afiican

cichlids based on years of experimentation; and have met with some success. However.

increases in suspected nutritional diseases observed in the late 1990s now warrant

scientific investigation of the nutritional requirements and feed management for these

fishes. A survey was taken in March 2000 (Appendix A) of the African cichlid farmers

in South Florida to define feeding practices in use by the industry at that time. Poor








growth, pigmentation, and reproduction rates (as well as poor health status) were the

most common complaints by the farmers. From this survey, two popular commercial

feeds used on farms were evaluated for the effects of growth performance, general health,

survival, and changes in liver composition on two species of juvenile African cichlids

(Chapter 2).

Based on farm observations (Appendix A) and the present requests of the farmers,

development of a complete feed for the intensive production of Afiican cichlids is

warranted. This feed should produce similar growth rates to the commercial feeds

currently in use; and provide adequate pigmentation and general good health. Although

the nutrient requirements of ornamental African cichlids have not been documented, all

fish require the same 40 nutrients regardless of where the nutrients are derived (Rust

2002). Based on information present for other species, the quantity of these nutrients can

be estimated for cichlids. This is necessary in order to formulate practical diets for these

fishes.

Protein constitutes about 30 to50% of a fish's diet and is usually the most costly

component of the manufactured feed. Therefore protein is usually the first nutrient to be

quantified in formulating diets for "new" species. In Chapter 3 the effects of varying the

protein level on growth performance and hepatocyte changes in P. socolofi are evaluated.

One of the main goals of dietary formulations is to supply the minimum protein required

for optimum growth with an appropriate balance of other nutrients to supply the required

energy. This balance is termed the ratio of digestible energy to digestible protein

(DE:DP) (Sargent et al. 2002). Although some protein can be used for energy. this can

be minimized by including nonprotein, energy nutrients in the diet. Protein is usually








more expensive than other sources of dietary energy, and it is not desirable to use excess

protein as a significant source of energy in fish diets. Lipids are the primary source of

energy in fish diets and have a major effect in sparing protein as an energy source in fish

(Sargent et al. 1989; Wilson 1989) although the limit to its effectiveness is not well

defined for fish. In the last 20 years, lipid levels in feed formulations have increased

especially in the salmon and trout industry (Hillestad et al. 1998). For these species, an

increase in lipid levels successfully improved fish growth; but for others such as juvenile

sunshine bass (Morone chrysops x M. saxatilis) (Gallagher 1996) or tilapia (Tilapia

area) (Winfree and Stickney 1981), excess lipids increase carcass and visceral fat and

lipid storage in the liver (which can affect product quality and health of the fish).

Because of the metabolic interactions among protein, lipid, and carbohydrate, simply

defining exact dietary lipid requirements by itself is not meaningful (Sargent et al. 2002).

The total energy in the diet is a more useful factor in determining the minimum protein to

be included in the formulation. In Chapter 4, growth performance and hepatocyte

changes are evaluated while varying the energy levels at fixed protein levels. Diets that

contain excess energy may reduce feed intake or increase fat deposition in the body:

while diets deficient in energy may result in protein being used to meet energy needs

rather than for protein synthesis.

After determining the dietary protein, lipid, and total energy level suitable for

production of juvenile African cichlids, the formulation of a practical diet became

possible. Chapter 5 discusses the pros and cons of purchasing commercially prepared

diets versus farm-made feeds. We also propose a model diet for the commercial

preparation of a pelleted feed.









In summary, the objectives of this dissertation were

* to evaluate the differences that might be found in nutrient (protein. lipid. and

carbohydrate) requirements among juvenile African cichlid species of different

trophic feeding habitats (Chapters 3 and 4)

* to determine the requirements for protein and energy for these young fishes

(Chapter 4)

* to formulate a cost-effective model diet (Chapter 5).

This diet should improve nutrient use, provide growth rates equivalent to or better

than growth rates of commercial feeds currently in use. decrease the level of liver and

carcass fat deposition, and maintain good water quality.














CHAPTER 2
EFFECTS OF TWO COMMERCIALLY PREPARED FEEDS ON GROWTH
PERFORMANCE AND LIVER COMPOSITION OF JUVENILE AFRICAN
CICHLIDS (Pseudotropheus socolofi AND Haplochromis ahli)


Introduction

African cichlids make up one of the largest segments of the ornamental fish

industry in Florida and are a significant cash crop for farmers. The farm-gate value of

individual fish ranges from $0.75 to $100, depending on species, and the annual

production from all the farms in South Florida is estimated to be approximately $10

million (Florida Tropical Fish Farms Association (FTFFA), personal

communication). The success of culturing more than 500 species of African cichlids

outside of their natural habitat is attributed to the ability of the larvae to eat

commercial feeds, such as flakes or crumbles, that are of a large particle size (250 to

400 p) at the initiation of exogenous feeding (Chapman 2000). Despite this success.

nutrition is currently perceived as a major limitation to production of African cichlids

in Florida (FTFFA, personal communication). The size of the ornamental fish

market, which is relatively small compared to that of the salmon (Oncorhvynchus

spp.), tilapia (Oreochromis spp.), and channel catfish (Ictalurus pznctatus) industries,

has limited research investments into nutrition of ornamental fish. Consequently.

most farmers rely on feeds developed for these other aquaculture industries. In their

natural habitat, various species of African cichlids exhibit feeding behavior that can

be categorized as herbivorous, carnivorous, or omnivorous (Fryer 1959). Ho\\ever.








under intense culture conditions, fish are not categorized by their trophic ecology and

are all fed the same diet. Although salmon and trout (i.e., carnivores) diets contain

sufficient protein (>45%) to meet the requirements for growth of African cichlid

juveniles (Ako and Tamaru 1999), the level of lipid in the diet (15 to 20%) may be

excessive for them, (especially the herbivorous and omnivorous species), as

evidenced by their fatty livers and excessive fat around the viscera at necropsy

[University of Florida Fish Diagnostic Laboratory (UFFDL). personal observation].

During August 1998 and June 1999, African cichlid farms in South Florida

experienced fish mortalities of up to 20 to 30 fish per day. Examination of moribund

or dead fish by the UFFDL, showed fatty infiltration; heavy vacuolation; and severe

necrosis of the liver, pancreas, and spleen. Other clinical anomalies included

recurrent bacterial infections, and the presence of macrophagous inflammation and

mulifocal granulomas on all internal organs of the fish that were giemsa-positive but

acid-fast negative. The granulomas that stained geimsa-positive were indicative of a

Cryptobia iubilans infection. Although observed, these infections were not

significant enough to be the cause of mortalities on the farm. Hence these chronic

pathological changes in the internal organs were suspected to be as a result of an

underlying nutritional imbalance. Although no analyses were performed on the feeds

used by the farmers at the time, it was suggested that they change their feed for one

that had a lower lipid content (<10%); and supplement it with a vitamin premix.

Clinical signs in the affected farms were resolved after implementation of these

recommendations.








Our study evaluated the effectiveness of two commercially prepared feeds on

growth and survival of two species of African cichlids currently in commercial

production, one an omnivore (Pseudotropheus socolofi) and one a carnivore

(Haplochromis ahli). We also investigated hepatosomatic index (ratio of liver weight

to body weight) and histopathology of the liver, to determine potential liver pathology

in juveniles given artificial feed with a high (17%) versus low (7%) lipid content. It

was hypothesized that the omnivore, P. socolofi. would show a higher incidence and

severity of liver pathology than the carnivore, ahli when fed the high-fat diet.


Materials and Methods

A 12-wk feeding trial with juvenile African cichlids was conducted at the

University of Florida's Department of Fisheries and Aquatic Sciences. Gainesville.

Florida. Two commercially formulated fish feeds commonly used by African cichlid

farmers in Florida were chosen for the feeding trial (Table 2-1). These feeds had slightly

different protein levels (50%. 45%) but large differences in lipid content (17% versus

7%), and were representative of different feed types (200 -t trout fry starter sinking pellet

and mixed flake, respectively). Fry were hatchery bred and obtained from Angels

Hatchery. Homestead, Florida. Fish were 4 weeks of age, had an average body weight of

0.06 g and averaged 1.5 cm in total length. Six replicates were allotted for each diet in a

completely randomized block design. Fish were then stocked in 24. 34-L aquaria at a

rate of 20 per aquarium and acclimated on the experimental diets for 2 weeks. Fish were

fed at 6% body weight (BW) by hand two times per day for six days per week. Uneaten

food was removed daily with a siphon. Diets were stored at -20 OC until dispensed.








The research system used consisted of each aquarium set up on a partially closed

recirculating system. The water was recirculated at 2.5 L/minute through a biofilter and

UV sterilizer to remove impurities and reduce ammonia. Fish were held on a 12 h

light: 12 h dark photoperiod. Dissolved oxygen and water temperature were measured

every morning before each feeding. Dissolved oxygen was maintained near saturation

and the temperature was held between 28 to 30 C using 300W heaters in the sump.

Total ammonia nitrogen (TAN) and pH were recorded once per week. Total ammonia

nitrogen was held between 0.02 to 0.05 mg/L, pIT was 8, nitrite < 0.1 mg/L and total

hardness and alkalinity were 205 mg/L and 85.5 mg/L CaCO3. respectively.

Proximate composition of the diets (Table 2-1) was analyzed by standard methods

for crude protein, ether extract (for lipid) and ash (AOAC 1990). Gross energy was

determined using an adiabatic bomb calorimeter (Parr instrument CO.. Molne IL).

Carbohydrate concentration was determined by difference [100% (%crude protein + %

moisture + %ether extract + %ash + %crude fiber)].

To measure growth, fish were batch weighed at the beginning of the experiment

and at 2-week intervals during the experiment. Fish were fasted for 24 hours prior to

weighing. Adjustment for the quantity of feed to be fed on a body weight basis was done

at this time. Fish were measured for total length at the beginning and the end of the trial,

and were counted every week in order to calculate the mean survival per treatment.

Growth was analyzed first as a repeated measures analysis of variance (ANOVA) in a

completely randomized bock design to detect the interactions among fish species, diets.

and weeks (SAS 2000). A second one-way ANOVA was then conducted to determine

the differences in mean growth among weeks for each treatment. Within each sampling








period, differences among means were tested using Duncan's Multiple Range Test

(Duncan 1955). Mean length and survival were also analyzed by a one-way ANOVA

and differences in the means analyzed using the Duncan's Multiple Range Test (Duncan

1955).

At the beginning of the feed trial, 10 fish from each species were euthanized by

over anesthetization in a water bath with Ig/L of buffered tricaine methane sulfonate

(MS-222) (Tricaine S, Western Chemical, Inc., Ferndale. Washington) to evaluate their

parasite load, which is used as an indicator of health status. Presence or absence of

parasites was assessed by taking mucus, skin, and gill biopsies (Noga 2000). After the

final weighing, at 12 wks, four fish from each treatment were euthanized as above in

order to observe histological changes in the liver and for hepatosomatic index (HSI)

analysis where HSI = 100*(wet liver weight/wet weight of the whole fish). All livers

were excised on ice to minimize any changes that may occur in cytoplasmic contents of

the hepatic tissue.

Livers were weighed to the nearest 0.001 g and then immediately fixed in 10%

neutral buffered formalin. These samples were subsequently submitted for histological

preparation and examination. After embedding in paraffin and thin sectioning at 8ptm,

two liver samples per treatment were mounted on glass slides and stained using standard

hematoxylin and eosin (H&E) staining techniques (Luna 1968). Stained sections were

examined under a compound microscope and compared to normal tissues in standard

texts (Yasutake and Wales 1983; Ferguson 1989: Roberts and Ellis 2001; Rust 2002).

After histological processing, areas occupied by lipid droplets appeared as clear circular

or polygonal areas. For each treatment, three fields of view were randomly selected from









which measurements of the area occupied by the lipid droplets were taken. Lipid filled

areas were recorded as the percent liver droplet area (LDA) and were measured using

Image-Pro computer software. In addition, pathological liver conditions were scored

on a semi-quantitative scale ranging from 0 to 4: (0) absent; (1) tissues having sparse

scattering of lipid droplets; (2) tissues having lipid droplets in moderate concentrations

throughout the section; (3) tissues having an abundant accumulation of lipid droplets: and

(4) tissues showing necrotic cells due to excess accumulation of lipid droplets.

Differences in HSI and LDA between species and diets were analyzed by a two-

way ANOVA with species and diet as factors. Multiple comparisons were made with the

Duncan's Multiple Range Test (Duncan 1955).

Results

At the initiation of the feeding trial, P. socolofi and H. ahli were not significantly

different in weight or length among treatments (Table 2-2). and were on average 0.06 g

and 1.5 cm. At the end of the 12-week feeding trial, the mean body weight attained by P.

socolofi (1.14 g) and H. ahli (1.03g) on Diet 1 (pelleted feed) were not significantly

different (P > 0.05) (Table 2-2, Figure 2-1), nor were their mean final lengths different

(both 4.8 cm) (Table 2-2). Pseudolropheus socolofi on Diet 2 (flake). however, was

significantly smaller both in weight and length than either species on Diet 1 (Figure 2-1).

By the end of the 12-wk trial, H. ahli on Diet 2 had gained significantly less weight than

P. socolofi on Diet 2 but was not significantly different in length (Table 2). Survival

rates were greater than 95% for both species (Table 2-2). Postmortem exams conducted

on these fish produced no evidence of disease therefore mortality was attributed to

handling after weighing, with H. ahli more sensitive than P. socolofi.









On gross examination, livers of fish in both treatments were a pale cream color

and friable. Numerous hydrophobic lipid droplets could be seen Lcalecscing on fresh

liver-squash preparations. Excised liver tissue floated in fixative solution, indicative of

accumulated lipids in the liver. Comparisons among the HSI of the species and

treatments indicated that P. socolofi on Diets 1 and 2 and H. ahli on Diet 1 tended to have

higher HSIs (3.76 to 3.87%) compared to H. ahli on Diet 2 (2.23%). but the HSIs were

variable and overall were not significantly different (Table 2-2). Lipid droplets extracted

during histological processing left clear zones within the cytoplasm (Figure 2-2), which

were measured by image analysis as LDA (Table 2-2). Liver droplet area averaged 64%

for P. socolofi on Diet 1 (Table 2-2), revealing evidence of diffuse hepatic lipidosis.

which was scored as '3' (severe vacuolation) based on the liver pathology scale.

Hepatocyte boundaries ofP.socolofi on Diet 1 were very swollen and not well defined.

Nuclei were shrunken, irregular in shape and located at the periphery of the cell rather

than in the typical central position (Figure 2-2A). Haplochromis ahli on Diet land P.

socolofi on Diet 2 had liver pathology scores of'2' (moderate vacuolation) and an

average LDA of 36% and 29%, respectively. The livers of these fish had more uniform

lipid deposition throughout the hepatocytes (Figure 2-2B and 2-2C). The livers of H ahli

on Diet 2 had distinctly less lipid droplets with an average LDA of 0.8%. and hepatocytes

were more defined with more central nuclei (Figure 2-2D). These fish had liver

pathology scores of '1' (mild vacuolation).

Discussion

Our results show that when juvenile African cichlids are fed a high-energy, lipid-

rich diet in captivity, they develop fatty livers. Fatty liver is diagnosed by the presence of

intracellular accumulation of lipids in the hepatocytes, enlargement, pale color, and









friability of the liver (Ferguson 1989; Penrith et al. 1994; Brusle and Anadon 1996).

Fatty infiltration of the liver has also been attributed as "the most common metabolic

disturbance in aquarium fish" (Penrith et al. 1994). Intracellular accumulation of lipids in

the liver also affects a wide variety of other species, such as channel catfish (Judd and

Cross 1966), flag cichlid Cichlasomafestivum (Cooper 1972), juvenile cod Gadus

morhua (Grant et al. 1998), juvenile sea bass Dicentrarchus labrax (Mosconi-Bac 1990),

captive African stonefish Synanceja verrucosa (Penrith et al. 1994). and red drum

Sciaenops ocellatus (Tucker et al. 1997). The common factor among these species

appeared to be the prolonged feeding of a high-energy. high-lipid (e.g.. > 10% lipid)

artificial diet. Spontaneous hemorrhaging, anemia, infertility and immune suppression

with secondary infections have been commonly associated with fatty infiltration of the

liver, viscera and other organs (Ferguson 1989). As seen in Paperna et al. (1977).

suppression of the immune system due to fatty infiltration of these vital organs is a

possible explanation for recurring bacterial infections and chronic Cryptobia-induced

granulomas in fish that were submitted to the UFFDL for necropsy in 1998 and 1999.

From our feeding trial, it was apparent that African cichlid juveniles are capable

of achieving rapid growth when fed a diet containing 52% crude protein (CP) and 17%

lipid (Diet 1) (Figure 2-1). However, our results also showed that when these young

African cichlids were fed this high-energy, lipid-rich diet in captivity, they developed

fatty livers. Although histological analysis of fish on Diet 1 showed an increase in

hepatocyte changes associated with lipid storage, no other pathological changes such as

necrosis or ceroid (by-products as a result of oxidation of lipids) deposition were evident.

nor was growth or survival affected (Figure 2-1, Table 2-2). As in our study. Tucker et








al. (1997) showed that although there were hepatocyte changes associated with lipid

storage, pathology associated with necrosis was not evident in red drum fatty livers.

Histological investigations showed similar effects of artificial feed on the livers of

juvenile sea bass (Mosconi-Bac 1987; Mosconi-Bac 1990). As in our study, the

accumulation of lipids in the liver was shown to be an early and rapidly developing

phenomenon and this increase in liver lipid (fatty liver) was associated with an increase

in dietary lipid.

The commercial ornamental fish farming industry has expressed concern about

the occurrence of the fatty liver condition in their fish, which may lead to mortalities.

Considered as a true nutritional pathology, fatty degeneration of the liver is a fatal

progressive disease that has been described in salmonids (Roberts and Ellis 2001) and

cichlids (Ferguson 1989). However, Mosconi-Bac (1990) believes that lipid deposition

in hepatocytes of fish eating artificial feeds is a short-term condition of the hepatocytes

due to metabolic changes in the liver. Paperna et al. (1977) showed that the hepatocytes

return to normal after feeding a balanced diet was resumed. Lipid is an important source

of non-protein dietary energy and, in general, an increase in lipid in the diet improves

protein utilization. However, excess lipid in the diet can result in an imbalance in the

digestible energy to crude protein ratio (DE:CP) (NRC 1993; Wilson 1995). This

imbalance may result in excessive fat deposition in the viscera and tissues (Wilson 1995)

and can lead to progressively degenerative changes in liver cells (Tucker et al. 1997).

Jauncey (2000) reported that tilapia, similar to P. socolofi in having herbivorous

and omnivorous feeding behavior, do not utilize high levels of dietary lipids as

effectively as salmonids. An increase of total lipid in the diet greater than 12% leads to a








deposition of large amounts of body fat and accumulation of fat in the liver of juvenile 0.

aureus x 0. niloticus hybrids (Jauncey and Ross 1982). As in the tilapia hybrids. P.

socolofi fed Diet 1 had excess (>50% of the cells) lipid accumulation in the hepatocyes

(Figure 2-2A). In order to maximize protein utilization and efficiency. Jauncey (2000)

suggested that optimal lipid levels in tropical feeds for tilapia should be between 6 to

12%. When lipid levels were reduced to 7% in the diet, as for P. socolofi consuming Diet

2, the percent lipid deposited was reduced (Figure 2-2C). For tilapias the DE:CP is

recommended to be 8.2 (56% protein and 4,600 kcal/kg digestible energy for tilapia fry

up to 2.5 g). In our study, (where it was assumed that due to the highly digestible

ingredients used, the DE approached the GE for the diets), the gross energy to crude

protein ratio (GE:CP) (Table 2-1) for Diet 1 was 8.0 (52% CP and 4190 kcal/kg) and for

Diet 2 was 7.5 (46.8% CP and 3520 kcal/kg).

The slower growth and reduced lipid reserves of H. ahli fed Diet 2 may have been

due to the low GE:CP ratio of the diet. Carnivorous species generally require higher

levels of protein and energy for growth and are therefore able to utilize higher levels of

lipids as a source of non-protein energy than omnivores (Smith 1989). Lipid levels

between 15 to 20% are included into the diet of carnivorous species such as trout (Wilson

1995; Wantanabe 1998). Haplochromis ahli is piscivorous in its natural habitat and Diet

2 may not have provided a sufficient source of non-protein energy needed to supply the

calories for rapid growth. These fish had <10% lipids deposited in the hepatocytes

(Figure 2-2D). In this situation, protein may have been used as an energy source instead

of for tissue synthesis, thereby causing a reduction in the growth of H. ahli. On the other

hand, H. ahli fed Diet 1 grew without excessive (<50%) accumulation of lipids in the









hepatocytes (Figure 2-2B) compared to P. socolofi, the omnivorous species, which had

excessive accumulation of lipids.

Ornamental fish feeds have traditionally been produced for the aquarium industry

using flake technology. Flakes are suitable for community fish in an aquarium tank as

they have a high surface area to volume ratio and sink slowly, allowing fish at different

feeding levels (top-, mid- and bottom feeders) to consume the flakes (Winfree 1992).

This large surface area, however, can result in the rapid leaching of water-soluble

nutrients, such as essential amino acids, vitamins C and B-complex (Barrows and Lellis

2000). The high temperatures required for drying flakes also denature proteins, oxidize

lipids and destroy vitamins and carotenoids, which are essential for the rapid growth of

fry (Barrows and Lellis 2000). Flakes are also usually designed as a maintenance diet

and not to optimize growth or reproduction (Ako and Tamaru 1999). Complete pelletized

feeds for food fish have been based on a small number of economically important

species, primarily salmon, trout and catfish. These feeds have the advantage of being

more nutrient dense and stable in water. They have also been designed to obtain

maximum growth rates at minimum costs for select species. When feeding marble

angelfish Plerophyllum scalare, Ako and Tamaru (1999) compared a flake diet used in

the aquarium industry with a pelleted feed designed for salmon fry. They found that

although the flake and salmon feed contained similar total amino acid levels, the salmon

feed contained higher levels of methionine, histidine and cystine compared to the flake

food. The flakes also contained lower levels of essential and total fatty acids. Therefore.

not only do salmon feeds have a higher percentage of protein needed for the growth of

young fish, but they contain all the essential nutrients as well as carotenoids needed for








ornamental fish. The major difference with these feeds compared to feed typical for

aquarium fish is their higher percent lipid content. The higher percentage of marine fish

oil in the feeds for salmonids has the advantage of acting as an attractant and is more

highly palatable to the fish. This increases their feed intake and allows faster growth

rates and higher yields. Ako and Tamaru (1999) therefore concluded that the salmon

feed tested was superior for growing marble angelfish at lower costs than the flake feed.

The present feeding trial demonstrated that a commercially prepared fish diet high

in protein and lipid fed to young African cichlids can produce fast growth rates for the

first 12 weeks. Based on observed growth rates, fish fed the nutrient-dense. pelleted feed

seems to be the more efficient way of feeding commercially farmed African cichlids.

The trade-off was the fact that this diet induced a fatty liver condition, indicating that the

lipid level in this diet was too high for African cichlid juveniles. With prolonged feeding

of a high-energy, lipid-rich diet, degenerative changes of the liver leading to death can

occur if the diet is not corrected (Ferguson 1989; Tucker et al. 1997). The lower-energy.

flaked diet might not be suitable for producing African cichlids on a commercial level,

especially carnivorous species, as fish on Diet 2 had lower growth rates. This diet.

however, might be suitable for maintenance of cichlids in a home aquarium, as the health

of the fish was not jeopardized.

Based on growth, survival and absence of obvious clinical disease. the fatty liver

condition observed in juvenile African cichlids in this feeding trial did not appear to have

any immediate adverse health effects. However, the deposition of excessive lipid stores

in the livers in juveniles fed the high energy feed, especially the omnivorous species P.

socolofi, indicated the need for future studies on the optimum energy to protein ratio in






27

the diet required for growth without deposition of excess lipid in the liver and viscera of

captive-reared African cichlids.











Table 2-1. Proximate composition of two commercial diets used
in feeding trials for juvenile African cichlids.


Nutritional composition
Dry Matter (%)
Crude protein (CP) (%)
Ether Extract (%)
Carbohydrate (%)
Ash (%)
Gross Energy (GE) (kcal/g)
GE:CP ratio (kcalGE /g CP)


Diet 1 (pellet)
92.6
52.0
17.0
14.5
9.1
4.2
8.0


Diet 2 (flake)
93.6
46.8
7.1
25.1
14.6
3.5
7.5










Table 2-2. Mean growth performance and hepatocyte changes of juvenile African cichlids fed
two commercial diets (high lipid pellet versus low lipid flake) for 12 weeks.


Survival (%)
Initial weight (g)
Final weight (g)
Initial length (cm)
Final length (cm)
Hepatosomatic Index (%)
Lipid Droplet Area (%)
Score2


Diet 1
P. socolofi
99.0a + 0.1
0.07a + 0.01
1.14a 0.03
1.50a 0.01
4.77a L 0.24
3.81a 0.53
64.1a 2.8
3


(pellet)
H. ahli
95.08" 0.6
0.06a 0.01
1.03a+ 0.05
1.50a 0.01
4.83" 0.17
3.76a + 0.26
36.5 2.3
2


Diet 2
P. socolofi
98.0a 0.3
0.06a 0.01
0.56b 0.02
1.50a 0.01
3.66' 0.17
3.87a 0.35
29.1b 2.0
2


(flake)
H. ahli
95.0 0.3
0.06a 0.01
0.44c + 0.02
1.50a 0.01
3.85b 0.06
2 23; 0.62
8.8c 4.0
1


'Values in each column with the same superscript are not significant different (P > 0.05).
Data are expressed as means SE. Mean weight n=6: Hepatosomatic index n=4: LDA n=3.
2Score = graded scale of pathological changes in the liver: 0=absent: 1=mild: 2=moderate:
3=severe; 4=necrosis and cell death


--















1.20



.00-- Diet I:P. socoloti

Dietl:H.ahli
--6- Diet2:P. socolofi
0.80
0.80 -- Diet2:H. ahli



D 0.60


E
0.40



0.20



0.00 ''
0 3 5 7 10 12

weeks

Figure 2-1. Mean weight of juvenile African cichlids (P. socolofi and H. ahli) fed
commercial Diet 1 (17% lipid; 4.2 kcal/g pellet)) or Diet 2 (7.1% lipid; 3.5 kcal/g flake)
for 12 weeks.











NL



1. 10 i


Figure 2-2. Histological sections from the livers of P. socolofi and H. ahli fed
experimental diets for 12 weeks: (A) P. socolofi on Diet 1 (pellet), representative of liver
pathology code '3', with hepatocytes distended with fat. Note small, hypochromatic
nuclei at periphery of cell (arrows); (B) H. ahli on Diet 1, representative of liver
pathology code '2', with moderately vacuolated liver tissue; (C) P. socolofi on Diet 2
(flake), also representative of code '2' with moderately vacuolated liver tissue; and (D)
H. ahli on Diet 2, representative of liver pathology code '1', with mildly vacuolated
tissue. Note that cells are more defined. L = lipid droplets













CHAPTER 3

EFFECTS OF DIETARY PROTEIN LEVELS ON GROWTH PERFORMANCE,
PROTEIN EFFICIENCY AND HEPATOCYTE CHANGES IN JUVENILE AFRICAN
CICHLIDS (Pseudotropheus socolofi)

Introduction

The cichlid fishes of Lake Malawi (Malawi) and Lake Tanganyika (Tanzania) are

prized for their intense coloration, vivacious behavior and for the vast array of species

present (Loiselle 1994; DeMason 1995). These fish are produced intensively in Florida

for the ornamental fish trade, with farm-gate prices ranging between $0.75 to $100 per

fish. Cichlids have the advantage over many species in having a functional stomach at

first feeding (Howell et al. 1998). Hence they are able to immediately assimilate artificial

diets containing high concentrations of easily digested nutrients. Presently, commercial

farmers use brine shrimp (Artemia salina) nauplii and Spirulina flakes at first exogenous

feeding, which is gradually replaced with artificial diets obtained from the food fish

industry (specifically salmon and trout) for grow-out of juveniles. These commercial

feeds have been shown to be useful in the production of ornamental cichlids but have

their limitations (Ako and Tamaru 1999; Chapter 2). Although these commercial feeds

are formulated for superior growth and reproduction rates, the unique qualities of

ornamental cichlids are not taken into consideration in the formulations; such as their

small size, intense pigmentation and their diverse feeding habits (e.g., herbivore versus

carnivore). The high lipid (15 to 20%) content in these feeds has also been shown to

cause fatty livers due to excess lipid in the diet of African cichlids that were fed a salmon








starter diet (52% crude protein and 17% lipid) (Chapter 2). These problems have

stimulated research in the development of a practical and cost-effective diet for farmed

African cichlids.

Feed contributes approximately 30 to 40% of the total operating costs of culturing

African cichlids in Florida (Florida Tropical Fish Farms Association [FTFFA], personal

communication). Of these costs, protein, which is essential for tissue growth, is the most

expensive dietary ingredient and represents more than 40% of the total cost of the feed

(Meyers 1998). Diets deficient in protein may reduce growth rate, while excess protein

increases feed costs as well as contributes to poorer water quality in the grow-out

facilities. Although African cichlids have been produced intensively in Florida for the

last 40 years, published studies on the growth and health performance of these fish are

lacking. However, protein requirements within the Cichlidae have been extensively

studied with Tilapia species because of their value as a food fish. Protein requirements

for tilapia fry to first feeding (0.5 g) ranges from 36 to 50% (Davis and Stickney 1978;

Jauncey and Ross 1982; El-Sayed and Teshima 1991), and for juvenile tilapia (0.5 to 5 g)

from 29 to 40% (Cruz and Laudencia 1977; Mazid et al. 1979; Jauncey 1982; Siddiqui et

al. 1988; De Silva et al. 1989). For the Central American cichlid Cichlasoma synspilum,

an omnivorous species, protein requirement was determined to be 40.81% (Olvera-Novoa

et al. 1996). The range in requirements observed in these studies are influenced by

several factors, such as daily feed allowance, fish stocking density, size of the fish, the

amount of non-protein energy in the feed and the quality of dietary protein (Steffens

1981; Wilson 2002).








The objective of this study was to determine the effects of varying the dietary

protein levels on growth performance, protein utilization, and hepatocyte changes of

juvenile African cichlids, using Pseudotropheus socolofi as a model omnivorous species.

P. socolofi is an appropriate African cichlid to use in experimental conditions because it

is cultured in Florida, it easily accepts artificial diets, and it is not as sensitive to handling

as other African cichlid species (Chapter 2).

Materials and Methods

Culture System and Experimental Design

Pseudotropheus. socolofi fry with yolk sac attached were obtained from Warner

Aquatic Resources Inc., Ruskin, Florida, and reared at the University of Florida,

Department of Fisheries and Aquatic Sciences Aquatic Facility. Fish were fed Spirulina

flakes (45% protein; 7% lipid) three times per day to satiation after depletion of their yolk

sac. At 2 weeks post-hatch (0.09 g, 1.8 cm) fish were randomly distributed into 18, 37-L

glass tanks so that each contained 20 fry per tank. Fry were acclimated to the

experimental system and test diets for 2 weeks. All diets were tested simultaneously with

three replicate tanks in a completely randomized design. Fish were fed at 12% body

weight (BW) per day by hand or by automatic feeders three times per day for six days per

week. Uneaten food and waste were removed daily with a siphon.

In the feed trial, each aquarium was part of a partially-closed recirculating system

that included a sump, biological filter and UV-sterilizer. Fish were held on a 12h light:

12h dark photoperiod. Dissolved oxygen and water temperature were measured using a

YSI Model 55 oxygen meter (YSI Industries, Yellow Springs, Ohio) every morning

before each feeding. Dissolved oxygen was maintained near saturation and the








temperature was held between 28 to 30 oC using 300W heaters in the sump. Total

ammonia nitrogen (TAN), nitrite and pH were recorded once per week. Total ammonia

nitrogen was held between 0.02 to 0.05 mg/L, nitrite < 0.1 mg/L, pH was 8, and total

hardness and alkalinity were 300 and 85.5 mg/L, respectively.

Experimental Diets

Six semi-purified diets were formulated as shown in Table 3-1. Dietary protein

was provided by a fixed level of menhaden fish meal (25%) and variable levels of

vitamin-free casein to produce six diets with analyzed protein content ranging from 32.52

to 58.86 %. Fish meal ensured adequate feed palatability and acceptance while providing

dietary amino acids. Wheat starch was used as a carbohydrate source and binding agent,

menhaden fish oil and soybean oil (1:1) as lipid sources, carboxymethylcellulose (CMC)

as the binder and cellulose as the filler. To prepare the diets, fish meal and wheat were

ground and passed through a 2-mm screen using a Wiley mill (Thomas-Wiley

Laboratory mill Model 4, Thomas Scientific, Philadelphia, USA). All dry ingredients

except choline chloride were weighed and mixed in a kitchen mixer (KitchenAid Model

KSM90, Greenville, Ohio) for 5 minutes. Choline chloride was then added and the

mixture was homogenized thoroughly for an additional 5 minutes before oils and boiling

water were added. Mixing was continued to give a total mixing time of 15 minutes for all

ingredients. A dough was formed and passed through a pasta maker (KitchenAid Model

KSM90, Greenville, Ohio) to form "spaghettis" 1-mm in diameter, which were then cut

with a kitchen knife into pellets approximately 3-mm in length. Pellets were dried in an

air-convection drier at 320C for 4 hours and sieved to achieve consistency in pellet size

prior to feeding. These diets were stored at -20 oC for long-term storage, but small








portions for weekly feeding were maintained in the refrigerator. Proximate analyses were

carried out in duplicate for moisture, crude protein, lipid and ash for major ingredients

and each formulated diet (Table 3-1). Feeds were ground in a coffee grinder and freeze-

dried to determine % moisture content. Protein was determined by the Kjeldahl method,

lipid by ether extraction, and ash by combustion at 600 oC for 16 h (AOAC 1990). Gross

energy was determined using a Parr 1261 isoperibol bomb calorimeter and corrected for

fuse wire bur and acid production.

Growth Performance

To measure growth, fish from each tank were batch-weighed at the beginning of

the experiment and at 2-week intervals during the experiment, which ran for a 10-week

period. Fish were fasted for 24 h prior to weighing. Adjustment for the amount of feed

to be fed on a body weight basis was done at this time. Fish were measured for total

length at the beginning from a random subsample of the pool of fish. At the end of the

trial, six fish from each diet were measured for total length. Fish from each tank were

counted every week in order to calculate the mean survival per treatment. Growth

performance was measured in terms of survival (%); final body weight (g); weight gain,

WG (g) = final weight initial weight; feed conversion ratio, FCR = feed intake / weight

gain; and final length (cm).

The effects of time on growth and dietary effects over time were analyzed using a

repeated measures analysis of variance (ANOVA) (SAS 2000). A second one-way

ANOVA was then conducted to determine the differences in mean growth among

treatments within each sampling period, and differences among means were tested using

a Student-Newman-Keuls (SNK) multiple range test (Zar 1996). Survival, final length,








final weight gain, and FCR were analyzed by one-way ANOVAs and SNK multiple

range tests were used to distinguish significant differences among treatment means for

each variable.

Protein Use and Requirement

An initial random sample of fish at the beginning of the experiment, and a sample

of fish from each treatment at the end of the experiment, were killed with 1 g-L'' of

buffered tricane methane sulphonate (MS-222) (Tricaine S, Western Chemical, Inc.,

Ferndale, Washington) for proximate analysis in order to evaluate body composition.

Fish were then freeze-dried to determine % moisture content and ground in a coffee

grinder. Protein was determined by the Kjeldahl method and lipid by ether extraction

(AOAC 1990). Protein utilization was determined by the protein conversion efficiency

(PCE) calculated as (dry protein gain* 100) / protein offered (Thoman et al. 1999).

Protein efficiency ratio (PER) was calculated as weight gain (g) / protein intake (g).

Protein conversion efficiency and PER were analyzed using one-way ANOVAs and,

when significant, differences among the treatment means were determined using a SNK

Multiple Range Test.

Hepatocyte Changes

To assess changes in the liver due to diet differences during the experiment, six

fish from each treatment were euthanised as above after the final weighing at 10 weeks.

Livers were then excised while the fish were on ice to minimize any changes that may

occur in cytoplasmic contents of the hepatic tissue. Livers for calculation of a

hepatosomatic index (HSI) were weighed to the nearest 0.001 g and HSI was calculated as

100*(wet weight of the liver / wet weight of the whole fish). The HSI was analyzed by a








one-way ANOVA and differences among the treatment means determined using a SNK

multiple range test. Tissue samples from the livers of one fish from each treatment were

also processed using frozen/cryostat sections stained with Oil red O (Carson 1990) and

examined under a compound microscope. After histological processing, areas occupied

by lipid droplets were stained red with Oil red O. The size of the lipid droplet area

(LDA) was measured using Image-Pro computer software, Silver springs, Maryland.

The LDA for an individual fish was calculated as the mean of the LDA observed in three,

randomly selected fields of view, and differences among treatments were compared

qualitatively.

Results

Diets were formulated to achieve protein levels between 30 to 55%, with an

isocaloric energy level of approximately 4000 kcal-kg-'. However, proximate analysis

revealed higher protein levels than that formulated (Table 3-1) and dietary protein levels

ranged from 32.5% to 58.9% and gross energy levels ranged from 4740 to 5150 kcal-kg1.

Mean growth of P. socolofi fed at the various dietary protein levels during the 10-

week feeding trial was positive, and had a graded response of fish growth versus protein

level (Figure 3-1, Table 3-2). At the initiation of the feeding trial, P. socolofi on all diets

were not significantly different in weight (P = 0.25) or length (P = 0.30) among

treatments, and were on average 0.11 g and 1.8 cm. At the end of the 10-week feeding

trial, the length offish was not significantly different among treatments (P = 0.21). Final

weight and final weight gain were significantly different among treatments (P = 0.04 and

P = 0.03, respectively), with both having a graded response. Fish on the 32.5% protein

diet had the lowest final weight (or weight gain), which was significantly different from








fish on a 42.5% diet, but not significantly different from fish fed a 36.9% diet or fish fed

diets ranging from 48.4 % to 58.9% protein (Figure 3-2, Table 3-2). One replicate (tank)

fed at 48.4% protein was not used in the analysis because of its outlying weight and

resulting heterogeneity that would otherwise be introduced into the analysis. No

statistical difference was observed in percent survival (P = 0.34), which ranged from

95.3% (Diet 5) to 99.3% (Diet 6) for the 10-week experimental period (Table 3-2).

The feed conversion ratio was high (2.75 to 3.35) for all diets and not

significantly different among treatments (P = 0.14) (Table 3-2). However, varying the

protein level resulted in significant differences in PER (P = 0.0001) and PCE (P =

0.0002) (Table 3-2). Both PER and PCE decreased as the protein content of the diet

increased from 32.5% to 58.9% (Figure 3-3), but did not significantly change between

treatments containing dietary protein levels of 36.9 to 52.0% (Table 3-2).

Initial and final body composition data are presented in Table 3-3. Dietary

protein content did not affect carcass composition with respect to moisture and protein

content (P = 0.10), however there was a significant difference in carcass lipid content (P

= 0.0001). Final carcass lipid composition of fish on diets of 36.9% and 42.5% was

significantly greater than the initial lipid composition, but not different from fish fed

32.5% and 48.4% protein. Overall, there was a graded response of lipid content with

protein fed, with lipid content lowest in fish fed diets of 52.0 % and 58.9% protein.

On gross examination, livers at the end of 10 weeks from all treatments were a

pale cream color. Numerous hydrophobic lipid droplets could be seen coalescing on

fresh liver-squash preparations. Excised liver tissue floated in fixative solution,

indicative of accumulated lipids in the liver (King and Alroy 1997). No pathology such








as necrotic hepatocytes was observed in any livers irrespective of treatment. There were

no significant differences (P = 0.46) among diets in HSI (2.7 to 3.9%) (Table 3-2). Lipid

Deposition Area was similar among all treatments and ranged from 40% to 59%.

Discussion

Based on experimental feeding trials using semi-purified diets of varying protein

levels, the growth performance of P. socolofi juveniles was related to the level of protein

in the diet over time (Figure 3-1). There was a graded response in mean final weight gain

with increasing dietary protein levels (Figure 3-2). Raising the protein level to 42.5%

produced greater final weight gain than fish fed a diet of 32.5%. However, similar to fish

growth on the 36.9% protein diet, growth of fish on diets of 48.4% to 58.9% protein was

intermediate to growth on 32.5% protein (low) and 42.5% protein (high) (Table 3-2,

Figure 3-2). This overall trend in growth response with increasing dietary protein levels

has been observed in other cichlids (Mazid et al. 1979; Jauncey 1982; Olvera-Novoa et

al. 1996). For example, weight gain of Tilapia zillii fingerlings increased with dietary

protein levels up to 34.7% and then decreased gradually with a further increase in protein

(up to 50% protein in the diet) (Mazid et al. 1979). The reduction in growth rate

observed in diets with excess protein has been postulated to be due to ammonia toxicity

(Zeitoun et al. 1976) or possibly due to inadequate non-protein energy necessary to

deaminate the high protein feed (Jauncey 1982). Although the optimal dietary energy

level for P. socolofi has not been determined, in the current experiment greater than 4000

kcal/kg of gross energy was provided for each diet, which was comparable to that used

for other cichlid species (Davis and Stickney 1978; Olvera-Novoa et al. 1996). None of








the diets fed induced pathological hepatocyte changes as determined by liver analyses

(HSI and LDA).

Protein efficiency ratio and PCE was higher at the lower protein level (32.5%)

(Figure 3-3), suggesting that at this level protein utilization for energy was reduced. In

contrast, protein was not being utilized solely for growth purposes when fed above 36.9%

to 42.5%. De Silva et al. (1989) reported a similar pattern for other cichlid species.

There was no effect of dietary protein level on the moisture or protein composition of

whole fish, which was also observed by Olvera-Novoa et al. (1996) in Cichlasoma

synspilum. Comparing the lipid content of fish before and after the experiment, the

carcass lipid content decreased with fish on diets containing the highest protein levels

(52.0% and 58.9%) protein. This trend was seen in Tilapia zilli fingerlings (Mazid et al.

1979) and also suggests that there were sufficient non-protein energy sources in the diet.

In the food fish industry, weight gain is usually considered as the most important

indicator of a marketable product (Lovell 1986). However, in the ornamental fish

industry, length is the parameter that ultimately determines the marketability of the fish.

In this study, the total final length was therefore of more value in determining growth

performance than weight gain. After 10 weeks on the experimental diets, the mean total

length of fish was not different among diets. However, based on the fish fed a diet of

32.5% protein having a trend towards both lower final weight and lower final lengths

(i.e., small and slender) (Table 3-2), this level of dietary protein would not be considered

adequate compared to fish fed the next highest protein levels, such as 36.9% to 42.5%

protein, which showed a graded increase in final lengths and final weights (Table 3-2).








Based on the growth performance, protein utilization and liver analyses, the

minimum range of protein that can be added to the diet of P. socolofi for optimal growth

can be obtained using diets containing 36.9% to 42.5% dietary protein (4820 to 4880

kcal/kg; 6.7% lipid). This range is comparable to what was determined for other juvenile

African cichlids (Davis and Stickney 1978; Mazid et al. 1979; Jauncey 1982; Siddiqui et

al.1988). Present diets used on commercial African cichlid farms utilize feeds

manufactured for trout and catfish production. This study shows that a protein

requirement of at least 36.9% in the diet for P. socolofi is lower than that for rainbow

trout (Satia 1974; Lovell 1998), but is similar to that for catfish (Lovell 1998). However,

low protein diets for catfish, generally contains higher levels of carbohydrates from plant

based ingredients (Lovell 1998) that are poorly utilized in some fish (Ferguson 1989);

while typical trout diets contain more fishmeal and/or other animal sources that are

higher in protein and fat than that of catfish diets (Lovell 1998). Presently there are no

published studies on levels of plant-based ingredients that can be added to the diet of

ornamental African cichlids. It is therefore recommended that diets that are formulated to

provide 36.9% crude protein in the diet of juvenile African cichlids should contain highly

digestible protein sources. Further studies are needed to investigate the effects of varying

the energy level (total lipid and carbohydrate) in the diet while controlling for the protein

level, in order to determine the minimum level of protein required that allows for optimal

growth and health of these fish. This will allow the development of a more cost-effective

feed, specifically formulated for African cichlids.











Table 3-1. Formulation and composition (g / 100 g dry weight) of experimental diets.
Diets
Ingredients 1 2 3 4 5 6
Fishmeala 25.0 25.0 25.0 25.0 25.0 25.0
Casein 14.0 19.0 24.0 30.0 35.0 40.5
Wheat starch 49.3 42.5 35.5 27.4 20.5 12.8
Fish oilb 2.5 2.5 2.5 2.5 2.5 2.5
Soybean oilb 2.5 2.5 2.5 2.5 2.5 2.5
Stay CC 0.4 0.4 0.4 0.4 0.4 0.4
Choline chloride 0.8 0.8 0.8 0.8 0.8 0.8
Vitamin premixd 1.0 1.0 1.0 1.0 1.0 1.0
Mineral premixe 1.0 1.0 1.0 1.0 1.0 1.0
Carboxymethylcellulose 3.0 3.0 3.0 3.0 3.0 3.0
Cellulose 0.5 2.3 4.3 6.4 8.3 10.5

Proximate Analysis
Moisture 8.6 6.6 6.6 6.1 4.0 4.8
Crude protein 32.5 36.9 42.5 48.4 52.0 58.9
Lipid 6.5 6.7 6.7 6.7 7.2 6.7
Ash 6.0 7.3 6.0 5.9 4.5 5.3
Gross energy (kcal/ kg) 4740 4820 4880 4920 5080 5150
Gross energy to protein ratio 14.6 13.1 11.5 10.2 9.8 8.8
aMenhaden (68% crude protein), Omega Protein, Hammond, LA
b Fish oil and soybean oil, Sigma-Aldrich Co.
c Stay C-35 (ascorbyl-2-polyphospahte) Hoffman-La Roche, Ltd.
d Federal Vitamin #30 by Zeigler Bros. Inc, Gardners, PA. To supply per kg of diet:
D-Calcium Pantothenate, 106 mg; Pyridoxine, 30.8 mg; Riboflavin, 52.8 mg;
Niacinamide, 220.4 mg; Folic Acid, 848 mg; Thiamin, 32.5 mg; Biotin, 0.35 mg; Vit B12,
0.22 mg; Menadione sodium bisulfite (Vit K), 11.4 mg; Vit E, 352.8 I.U.; Vit. D3, 441
I.U.; Vit. A, 6,615 I.U.
e Mineral premix #3 by Zeigler Bros. Inc., Gardners, PA. To supply per kg of diet:
ZnSO4, 165 mg; MnSO4, 27.5 mg; CuSO4, 5.5 mg; FeSO4, 48.4 mg
All other ingredients from ICN Biomedical Inc







44

M a.,

(5 0


f 0 O L N '-'
M0 0 00 0 00


40 00 -0



M C ofs s 0
a0 0 r00 C 0













v v v o o ^
6\0 I- Cl





00 0
4c o oj oc


0 0 Z ,g
m 0 m m a
Oi 00 00 W 0 c \'O



_~ C 00-'- 0 0 (





S- t- 0 0
0 0 0 C0& CI .
I 0








to m o2
m) 00ci 00 -m








t .S 00 l
.o Oo6 6o o oo

C CuD CD en w







000 N N 00 '
SN00 u .0 r000








N 4 4 4),
V 0mm
CI)mn .- .
-, _- C.)mo
C~~~~ ~-c00 4)a~~
*\OC ~c~ d v 4
4) dd

~ 3
z ^d -C
H '








Table 3-3. Carcass composition of P. socolofi after 10 weeks of feeding diets with
increasing protein content. Means ( 1SE, n=2) within a column with different
superscripts are significantly different (P < 0.05).

Body Composition
Diet Moisture (%) Crude protein' (%) Lipid' (%)
Initial 72.66 a (0) 55.67 a (1.12) 20.43b (1.01)
1(32.5)2 70.77 a (0) 55.39 a (1.21) 22.88a"b (0.01)
2 (36.9) 71.73a (0) 52.42 a (1.95) 23.78a (1.45)
3(42.5) 68.94 a (0) 52.58a (0.31) 25.62a (0.91)
4(48.4) 72.04a (0) 54.34a (1.35) 23.21a.b (0.49)
5 (52.0) 71.00 a (0) 53.32a (0.93) 20.09b-c (0.93)
6 (58.9) 72.64a (0) 58.55 a (1.74) 18.02b.c (1.74)
'Determined on a dry weight basis
2Crude protein level is shown in parentheses













1.298 32.5% CP

1.098
--i 36.9% CP
s -9-+- 42.5% CP
S0.898 -X- 48.4% CP

0.698 --o- 52.0% CP
.0 -- 58.9% CP
" 0.498

0.298

0.098
0 2 4 6 8 10
Time (week)

Figure 3-1. Mean growth response of P. socolofi fed experimental diets
with different dietary protein levels (%) for 10 weeks.









1.40

1.20

1.00

0 0.80

S0.60

" 0.40
iT
0.20

0.00
3


0


Dietary Protein Level (%)




Figure3-2. Mean final weight gain ( ISE) ofP. socolofi fed diets with varying dietary
crude protein levels (32.5 to 58.9%) for 10 weeks.



























35 40 45 50 55


1.4

1.2
o
1.0 a,

0.8

0.6 w

0.4 |
a.
0.2

0.0


Dietary Protein Level (%)



Figure 3-3. Protein conversion efficiency (PCE) and protein efficiency ratio (PER) of P
socolofi fed experimental diets with 32.5 to 58.9% crude protein for 10 weeks.













CHAPTER 4
EFFECTS OF VARYING DIETARY PROTEIN AND ENERGY LEVELS ON
GROWTH PERFORMANCE AND HEPATOCYTE CHANGES IN JUVENILE
AFRICAN CICHLIDS (Pseudotropheus socolofi AND Haplochromis ahli)

Introduction

In their natural habitat, African cichlids show variations in feeding behavior,

ranging from strict herbivory to omnivory to carnivory. Pseudotropheus socolofi and

Haplochromis ahli are representative of popular species of African cichlids produced

commercially in South Florida for the aquarium trade and are classified as omnivorous

and carnivorous respectively (Loiselle 1994; Smith 2000). A general relationship

between natural feeding habits and dietary protein requirements has been noted where

herbivorous and omnivorous species require lower concentrations of dietary protein while

carnivorous species require higher concentrations. This is primarily due to herbivores

and omnivores being able to utilize carbohydrates for energy more efficiently than

carnivores, while carnivores are able to utilize lipids for energy more efficiently than

omnivores (NRC 1993; Gatlin III 2001). To date, most herbivorous and omnivorous fish

require crude protein (CP) at 25 to 35% of the diet, whereas carnivorous species may

require CP at 40 to 50% (NRC 1993; Gatlin III 2001). Growth studies conducted with

juvenile P. socolofi (Chapter 3), an omnivorous species, indicated that it required at least

36.9% CP for maximized growth.

Protein constitutes the largest portion of the cost of feeds and therefore protein in

the diet should be utilized for growth and other vital functions rather than serve as an

energy source for fish. For juvenile P. socolofi, there was no increase in growth above








36.9% protein in the diet (Chapter 3). Many studies have shown that the protein content

in the diet can be reduced without decreasing growth if the caloric content of the diet is

increased by manipulating the level of non-protein ingredients, such as carbohydrates and

lipids (Winfree and Stickney 1981; El-Sayed and Teshima 1991). Although lipids and

carbohydrates in the diet can minimize the use of protein as an energy source, excess

lipids and total energy in the diet can produce fatty livers in fish (Ferguson 1989). Hence

a proper balance of protein and non-protein energy is needed to supply nutrients for rapid

growth and feed utilization while maintaining fish health (Winfree and Stickney 1981).

The objectives of this study were to determine: (1) the minimum level of protein

at the optimum energy level necessary to achieve suitable growth rates, specifically by

varying the protein and energy levels without ensuing pathological hepatocyte changes

(such as cell necrosis) for juvenile African cichlids; and (2) if there are differences in

these requirements for an omnivorous species (P. socolofi) versus a carnivorous species

(H. ahli). It was hypothesized that the omnivore, P. socolofi, would have a lower protein

requirement and show a higher incidence and severity of liver pathology when fed a high-

fat diet than the carnivore, H ahli.

Materials and Methods

Experimental Design and Culture System

An 8-week feeding trial with juvenile African cichlids was conducted at the

University of Florida's Department of Fisheries and Aquatic Sciences, Gainesville,

Florida. Based on previous trials (Chapters 2 and 3), P. socolofi and H. ahli were chosen

as model species to represent African cichlids produced in Florida. Juvenile fish were

hatchery bred and obtained from Paramount Farms, Vero Beach, Florida. Fish were








randomly distributed in 36, 37-L tanks so that 18 tanks contained P. socolofi (average

body weight 0.89 g, 4.1 cm in total length) stocked at a rate of 12 fish per tank and 18

tanks containing H. ahli (1.1 g, 5.7 cm) stocked at a rate of 7 fish per tank. Stocking rate

was based on the practice in the hobby of stocking at least 2.5 cm of fish per 3.7 L of

water in the aquarium. Fish were acclimated to their respective diets and to the

experimental system 2 weeks before sampling and data collection began. All diets were

tested simultaneously with three replicates in a completely randomized design. Fish were

fed at 10% body weight (BW)/day for the first 3 weeks and then at 6% BW/day for the

remaining five week period. Feed was distributed three times per day by hand or by

automatic feeders for six days per week. Uneaten food and waste were removed daily

with a siphon.

The research system consisted of each tank set up on a partially-closed

recirculating system that included a sump, biological filter and UV-sterilizer. Fish were

held on a 12h light: 12h dark photoperiod. Dissolved oxygen and water temperature were

measured using a YSI Model 55 oxygen meter (YSI Industries, Yellow Springs, Ohio.

USA) every morning before each feeding. Dissolved oxygen was maintained near

saturation and the temperature was held between 28 to 300C using two 300W heaters in

the sump. Total ammonia nitrogen (TAN), nitrite and pH were measured once per week.

Total ammonia nitrogen was held between 0.02 to 0.05 mg/L, nitrite < 0.1 mg/L, pH was

8, and total hardness and alkalinity were 300 and 85.5 mg/L, respectively.

Experimental Diets

Six semi-purified diets were formulated as shown in Table 4-1. Dietary protein

was provided by a fixed level of menhaden fish meal (25%) and vitamin-free casein,









which was adjusted to achieve diets with two levels of protein (formulated as 30% and

50% CP). Fish meal ensured adequate feed palatability and acceptance while providing

dietary amino acids. Wheat starch was used as a carbohydrate source and binding agent,

menhaden fish oil and soybean oil (1:1) as lipid sources, carboxymethylcellulose (CMC)

as the binder and cellulose as filler. To vary the dietary energy content at each specific

protein level, diets were formulated in 5% increments of lipid from 5 to 15% of the diet.

The percent of wheat starch in the diet was fixed at 20% for diets with 50% protein and at

40% for diets with 30% protein. This gave a total of six diets with a range of gross

energy values from 4,000 kcal/kg to 5,300 kcal/kg.

To prepare the diets, fish meal and wheat were ground and passed through a 2-

mm screen using a Wiley mill (Thomas-Wiley Laboratory mill Model 4, Thomas

Scientific, Philadelphia, USA). All dry ingredients except choline chloride were weighed

and mixed in a kitchen mixer (KitchenAid Model KSM90, Greenville, Ohio) for 5

minutes. Choline chloride was then added and the mixture was homogenized thoroughly

for an additional 5 minutes before oils and boiling water were added. Mixing was

continued to give a total mixing time of 15 minutes for all ingredients. A dough was

formed and diets were passed through a pasta maker (KitchenAid Model KSM90,

Greenville, Ohio) to form "spaghettis" 1-mm in diameter, which were then cut with a

kitchen knife into pellets approximately 3-mm in length. Pellets were dried in an air-

convection drier at 320C for 4 hours and sieved to achieve consistency in pellet size prior

to feeding. These diets were stored at -20 OC for long-term storage, but smaller portions

for weekly feeding were maintained in the refrigerator. Proximate analyses were carried

out in duplicate for moisture, crude protein, lipid and ash for major ingredients and each








formulated diet (Table 4-1). Feeds were ground in a coffee grinder and freeze-dried to

determine % moisture content. Protein was determined by the Kjeldahl method, lipid by

ether extraction, and ash by combustion at 600 OC for 16 h (AOAC 1990). Gross energy

was determined using a Parr 1261 isoperibol bomb calorimeter and corrected for fuse

wire burn and acid production.

Growth Performance

To measure growth, fish were batch-weighed at the beginning of the experiment

and at 2-week intervals during the experiment. Fish were fasted for 24 hours prior to

weighing. Adjustment for the amount of feed to be fed on a body weight basis was done

at this time. Fish were measured for total length at the beginning and the end of the trial,

and were counted every week in order to calculate the mean survival per treatment.

Growth performance was measured in terms of: survival (%); final body weight

(g); weight gain, WG (g) = final weight initial weight; percent weight gain, WG (%)

100*(final weight initial weight / initial weight); feed intake, FI = average feed

consumed per fish per day; feed conversion ratio, FCR = feed intake / weight gain; and

final length (cm). Each species was analyzed separately for each of the following

analyses. Effects of time on growth and dietary effects over time were examined for each

species using a repeated measures analysis of variance (ANOVA) (SAS 2000). A second

one-way ANOVA was then conducted to determine the differences in mean growth

among weeks for each treatment. Within each sampling period, differences among

means were tested using Duncan's Multiple Range Test (Duncan 1955). Survival, FI,

FCR, final length and final weight gains were analyzed by a one-way ANOVA and








Duncan's Multiple Range Test (Duncan 1955) was used to distinguish significant

differences among treatment means at the completion of the experiment (8 weeks).

Whole Body Composition

At the beginning and at the end of the experiment, 9 fish per diet from each

species were euthanized by over-anaesthetization in a water bath with 1 g/L of buffered

tricaine methane sulfonate (MS-222) (Tricaine S, Western Chemical, Inc., Ferndale,

Washington) for proximate analysis of body composition. Fish were then freeze-dried

and ground in a coffee grinder to determine % moisture content. Protein was determined

by Kjeldahl method and lipid by ether extraction (AOAC 1990). Protein utilization was

determined by the protein conversion efficiency (PCE) calculated as (dry protein

gain* 100) / dry protein offered (Thoman et al. 1999).

Hepatosomatic Index and Liver Condition

After the final weighing at 8 weeks, six P. socolofi and four H. ahli from each

treatment were euthanized as above in order to observe histological changes in the liver

and for hepatosomatic index (HSI) analysis where HSI = 100*(wet liver weight/wet

weight of the whole fish). Livers were excised while the fish were on ice to minimize

any changes that may occur in cytoplasmic contents of the hepatic tissue. Livers for HSI

were weighed to the nearest 0.001g and then immediately fixed in 10% neutral buffered

formalin. These samples were subsequently submitted for histological preparation and

examination. Two liver samples per treatment were embedded in paraffin and thin

sectioned at 8pm, mounted on glass slides and stained using standard hematoxylin and

eosin (H&E) (Luna 1968). After histological processing, areas occupied by lipid droplets

appear as clear circular or polygonal vacuolated areas. Stained liver sections were








examined under a compound microscope and compared to lipid-filled liver sections

reported previously for African cichlids (Ferguson 1989) and to normal liver tissue of

other species of fish (Yasutake and Wales 1983; Roberts and Ellis 2001). For each

treatment, the size of the vacuoles present in the hepatocytes was determined by selecting

three fields of view randomly from each slide and measuring the area occupied by the

lipid droplets using Image-Pro computer software (Silver springs, Maryland). Lipid-

vacuolated areas were recorded as the percent liver droplet area (LDA). In addition,

pathological liver conditions were graded on a semi-quantitative scale ranging from 1 to

4: (1) tissues containing 0 to 25% vacuolation; (2) tissues having sparse scattering of

vacuoles (25 to 50 % vacuolation); (3) tissues having lipid vacuoles in moderate

concentrations throughout the section (50 to 75%); and (4) tissues showing necrotic cells

due to excess accumulation of lipid (75 to 100%). Hepatosomatic index and LDA were

each analyzed using a 2-way ANOVA (SAS 2000) with species and diet as factors.

Differences between the treatment means were determined using Duncan's Multiple

Range Test (Duncan 1955).

To determine possible glycogen accumulation in the hepatocyes, liver samples

from each treatment were treated with human saliva (enzyme treated). Enzyme-treated

slides and untreated slides were stained with periodic acid-Schiff (PAS) reagent (Luna

1968). Glycogen remaining in the untreated liver sections stains red to magenta with the

PAS reagent (Post 1987).

Results

Experimental diets were formulated to achieve three lipid levels (5, 10 and 15%)

at two protein levels (30% and 50%). However, proximate analysis revealed values








higher than the formulated diets (Table 4-1). Values ranged from 35.4% to 56.6% crude

protein, 8.3% to 20.1% lipid and 4,000 to 5,300 kcal/kg gross energy.

Growth Performance

P. socolofi. An evaluation of the final weight ofP. socolofi fed the various diets

during the 8-week feeding trial revealed a significant difference among dietary treatments

over time (P = 0.0001) (Figure 4-1, Table 4-2). At the beginning of the feeding trial P.

socolofi were not significantly different in weight (P = 0.27) or length (P = 0.20) among

treatments (Table 4-2) and were on average 0.89 g and 4.1 cm. By week four, differences

in mean body weight among diets began to appear (Figure 4-1) and continued to increase

over time. At the end of the 8-week feeding trial, final body weight (P = 0.0003), weight

gain (P = 0.0002) and percent weight gain (P = 0.0007) were significantly greater for fish

on Diet 1 (55.0% CP; 19.4% lipid). Intermediate final weight gains were observed for

fish on Diets 2 (56.58% CP; 13.0% lipid), 3 (56.24% CP; 8.32% lipid), 4 (35.95% CP;

20.1% lipid), and 6 (35.6% CP; 9.5% lipid). Although fish on Diet 5 (35% CP; 11%

lipid) achieved lower weight gains, it was not significantly different than weight gains

with Diet 6. It was interesting to note that at 4 weeks into the experiment, P. socolofi on

Diet 1 were noticeably gravid and by the 8h week were holding fry. Hence the weight

increase that was observed after 4 weeks was partially attributed to the weight of the

eggs.

Feed was totally consumed by fish on all diets except for fish on Diet 5. There

were significant differences in FI (P = 0.02) and FCR (P = 0.009) among diets. Fish on

Diets 1, 2, and 4 had a similar and higher feed intake compared to fish on the other diets.








Fish on Diet 1 had a lower FCR (2.35) compared to fish on Diets 3 (2.92), 5 (3.89) and 6

(3.04) but was similar to that of fish on Diets 2 and 4.

Final length was similar for P. socolofi fed Diets 1, 2 and 6 and greater than fish

fed diets 4 and 5 (Table 4-2). Final length observed for fish on Diet 3 had a value

intermediate to these diets. No statistical differences were observed in percent survival

(P = 0.34), which ranged from 99.3% (Diet 6) to 95.3% (Diet 5) for the 8-week

experimental period (Table 4-2).

H. ahli. An assessment of final body weight and weight gain of H ahli over the

8-week feeding trial revealed no significant differences among dietary treatments. At the

initiation of the feeding trial, H. ahli on all treatments were not significantly different in

weight (P = 0.56) or length (P = 0.1) (Table 4-3) and were on average 2.04 g and 5.65

cm. By week 4 differences in body weight (P = 0.02) were apparent (Figure 4-2), but by

week 6 no statistical differences were found in mean body weight for H. ahli among the

diets (P = 0.18). This trend continued till the end of the 8-week feeding trial at which

time there were no significant differences in final body weight (P = 0.18), final weight

gain (P = 0.22), percent weight gain (P = 0.29) and total length (P = 0.774) among

treatments. Like P. socolofi on Diet 1, 6 weeks into the experiment H. ahli on Diet 1

became noticeably gravid. Survival was 100% for fish on all treatments. While the FI

was significantly different among dietary treatments (P = 0.04), FCR was not (P = 0.46).

Feed intake per fish was greatest for fish on Diet 1 (9.04 g/fish), lowest on Diets 2, 3, 5,

and intermediate on Diets 4 and 6 (Table 4-3).








Whole Body Composition

Initial and final body composition data and protein conversion efficiency after

feeding experimental diets for 8 weeks are reported in Table 4-4. Trends in body

moisture, protein and lipid for whole body composition were evaluated for trends and

were not statistically analyzed because not enough fish were available at the end of the

experiment from each replicate to conduct proximate analyses in duplicate. After feeding

the diets for 8 weeks, body moisture from both species on all diets did not appear to vary

greatly from the initial body moisture. Protein content of P. socolofi was lower after 8

weeks on the experimental diets compared with their initial protein content, while protein

content in H. ahli at 8 weeks did not appear to be different from their initial condition.

For both species, at all protein levels, as the dietary lipid levels increased in the diet the

percent body protein tended to decrease, with fish on Diet 3 (56% CP, 8% lipid) having a

higher percent body protein compared to fish on the other diets (Table 4-4). Also, for

both species, whole body lipid content increased when the dietary lipid content increased,

with P. socolofi having a higher percentage of body fat compared to H. ahli. Fish on

Diets 1 (55% CP, 19% lipid) and 4 (35% CP, 20% lipid) had a higher percent body lipid

composition than fish on the other diets.

Varying the protein level resulted in significant differences in PCE (P = 0.001) for

both species. Fish on Diets 4 and 6 (36% CP) had significantly higher PCE than fish on

the other diets. In addition, PCE for P. socolofi was always greater than PCE for H. ahli

on the same diet.








Hepatosomatic Index and Liver Condition

On gross examination, excessive fat could be seen covering the visceral organs of

both species on Diets 1 and 4 (19% and 20% dietary lipid, respectively). Also, the eggs

of both species on Diet 1 were extensively covered with fat. P. socolofi on Diet 2 had

some visceral fat while both species on Diets 3 and 6 had very little visceral fat. Fish on

Diet 5 were thin and emaciated with no visual fat.

Fish on all diets stored lipid in the liver, and livers from both species on all

treatments at the end of 8 weeks were a pale cream color. Excised liver tissue floated in

fixative solution, which was indicative of accumulated lipids in the liver. Numerous

hydrophobic lipid droplets could also be seen coalescing on fresh liver-squash

preparations. The livers of fish on the high fat diets (Diets 1 and 4) were enlarged,

extremely friable and broke apart on exposure during the dissection. Although fish on

Diet 5 were thin, the liver was enlarged, friable and liquefied when exposed on

dissection. Livers from fish on Diets 3 and 6 were firmer than the livers of fish on the

other diets.

Hepatosomatic indices between species and among treatments were significantly

different (P = 0.05) and indicated that P. socolofi tended to have a higher HSI (2.46% to

4.70%) compared to H ahli (2.47% to 3.07%) (Table 4-5). However, the HSI for H. ahli

was not significantly different among diet treatments (P = 0.98) while it was different for

P. socolofi (P = 0.05). The HSI for P. socolofi on Diet 5 was greater than fish on Diets 3

and 6, while HSI of fish on Diets 1, 2 and 4 were intermediate to these diets (Table 4-5).

Comparisons between species revealed that P. socolofi also had higher LDA

(55.81% to 73.70%) compared to H. ahli (28.22% to 55.56%) (Table 4-5). For P.








socolofi, fish on Diets 1, 4 and 5 had the highest percentages of LDA (64.6 to 73.7%)

whereas fish on Diets 2, 3, and 6 had significantly lower LDA (55.8 to 60.9%) (Table 4-

5). LDA for H. ahli on Diets 2 and 5 (52.8 to 55.6%) were significantly greater than fish

fed Diets 3 and 6 (28.2% and 39.1%, respectively), and Diets 1 and 4 had intermediate

LDA. Scores of mainly 3 (severe) and 2 (moderate) were assigned based on the liver

pathology scale to P. socolofi and H ahli, respectively. Hepatocyte boundaries of both

species on all diets were round or polygonal in shape (Figures 4-3 and 4-4). The nuclei

of both species on Diets 1, 4 and 5 were not well defined, while the nuclei of fish on the

other diets were shrunken and located at the periphery of the cell rather than in the typical

central position.

In the vacuoles ofH. ahli hepatocyte tissue on Diet 5 (Figure 4-5B), glycogen

granules were stained pink with PAS reagent. A few glycogen granules were also

observed around the periphery ofhepatocytes ofP. socolofi on Diet 5 (Figure 4-5A) and

H. ahli on Diet 6 (Figure 4-5D). Liver tissue of fish on all other diets did not contain

glycogen granules.

Discussion

In general, all diets were similar in consistency except for Diet 5, which had a

hard "rubber"-like consistency and seemed to be the least palatable diet. Both species

responded similarly to all the diets with the exception of visible bloating observed in H.

ahli on Diet 5. Both species did not consume Diet 5 readily and it was suspected that the

carbohydrate concentration (54%) and the hard 'rubber-like' texture of this diet

contributed to its unpalatability. Hard feed is difficult to digest and can lead to an

inefficient uptake of nutrients (FAO 1980) and slower growth. Diets 1 and 4 were the








most palatable as fish from both species consumed this diet within 5 minutes of feeding.

Diet 4, however, had a negative effect on water quality. Within a few minutes of feeding

Diet 4, lipid droplets could be seen floating on the surface and the water became cloudy.

Both species utilized in this study adapted favorably to experimental conditions as shown

by the high survival rates.

Diet 1 (55% CP, 19% lipid), which is similar in nutrient analysis (52% CP, 17%

lipid; Chapter 2) to diets presently being used on farms in Florida, provided rapid growth

for P. socolofi. These fish had the highest final weights and lengths and lower FCR

compared to P. socolofi on the other diets. Statistical differences for growth parameters

measured were not observed for H. ahli on all dietary treatments, however, both species

on this diet became gravid relatively early in the experiment. It is suspected that the high

nutrient content of the diet promoted early sexual maturation. Although growth

performance was enhanced for fish on Diet 1, on dissection gross visceral fat was

visually observed surrounding the internal organs and eggs of both species. Histology

revealed severe vacuolation (liver pathology score of 3) of the livers of P. socolofi on

Diet 1 and a moderate amount of vacuolation (score = 2) for H. ahli on the same diet.

Despite this, survival was 97% and 100% for P.socolofi and H. ahli, respectively, and

fish showed no external clinical pathology.

Based on the previous experiment with P. socolofi (Chapter 3), adding more than

36.9% CP in the diet provided no additional growth. The excess protein is therefore

being used for other metabolic processes. Having more protein than needed in the diet is

not cost effective and the increase in amino acid deamination increases the amount of

nitrogen excretion (Bureau et al. 2002). Although Diets 4 and 6 (36% CP) had less








protein than Diets 2 and 3 (55% CP), both Diets 4 and 6 had similar increases in weight

gains to Diets 2 and 3. Diet 6 also had a similar final length to that of Diets 2 and 3. This

indicates that diets providing 55% crude protein are in excess of what is required to

maintain growth rates in juvenile African cichlids, irrespective of the dietary lipid level.

Therefore, 36% CP in the diet would be utilized for tissue growth more efficiently, and

would also be more cost effective than the high protein diet. The results of this feeding

trial demonstrated that a high protein/low energy diet (Diet 3) or a low protein/low

energy diet (Diet 6) could promote acceptable growth in juvenile African cichlids. Based

on the PER, however, the protein in Diet 6 was utilized more efficiently than that in Diet

3 (Table 4-4). A greater percentage of dietary protein was therefore utilized for body

protein synthesis with the lower protein diet (Diet 6) than the high protein diet (Diet 3).

The type of non-protein energy added to the diet exerts a significant effect on

protein efficiency and utilization. Ogino et al. (1976) showed that there was no

difference in the protein requirement to get the maximum growth rate between rainbow

trout (Oncorhynchus mykiss) (a carnivorous species) and carp (Cyprinus carpio) (an

omnivorous species) if fish were fed a diet having the proper energy sources and levels.

Lipids are an excellent source of energy and on a per weight basis produces 2.25 times

more energy than either protein or carbohydrates (Robinson and Wilson 1985; Barrows

and Hardy 2001). Lipids also have a major protein-sparing effect in many species

(Sargent et al. 1989; Wilson 1989). However, feeding high lipid diets may cause fatty

infiltration of the liver and excessive obesity (Stoskopf 1992). Obese fish will have

distended abdomens with masses of adipose fat, which may limit the visibility of the

abdominal organs (Stoskopf 1992). This was observed in the African cichlids fed both








Diets 1 and 4. The lower crude protein in Diets 4 (36% CP, 20% fat) and 6 (36% CP,

9.5% fat) produced similar growth responses in both species of African cichlids.

However, fish on Diet 4 had more visceral fat and more fat was deposited in the liver

than fish fed Diet 6. This is not a desired effect. As suggested by Post (1987), obesity

can lead to liver malfunction, which can alter kidney function and lead to retention of

body water and edema of various organs, including the liver. Although not observed in

this experiment, it has been shown that prolonged increased vacuolation in the liver due

to excessive lipid storage can lead to necrosis of the liver and eventual death (Ferguson

1989; Roberts 2002). It has also been suggested that lipogenesis could be inhibited with

diets in excess of 10% lipid (Sargent et al. 1989). Based on the intracellular fat droplets

in liver tissue of P. socolofi that were observed microscopically in frozen sections stained

with Oil red O in Chapter 3, it was assumed that the vacuoles observed by H&E staining

in this study were lipid-filled vacoules. It was hypothesized that H. ahli (the carnivorous

species) should be able to tolerate higher dietary lipid than P. socolofi (the omnivorous

species). This study showed that when both species consumed greater than 10% dietary

lipid, excess lipid was stored in the liver. However, H. ahli on all diets stored less fat in

the liver (i.e., lower LDA and liver condition scores, Table 4-5) compared to P. socolofi,

suggesting that H. ahli was able to metabolize a higher lipid diet. Fatty infiltration of the

liver has also been reported for other warm water species, such as sunshine bass (Morone

chrysops x M. saxatilis), fed a diet containing 16% lipid (Gallagher 1996).

Ogino et al. (1976) determined that with diets for rainbow trout in which lipid was

used as the main energy source, protein utilization was high at low protein levels,

whereas when carbohydrate was used as the main energy source the protein utilization








was low at low protein levels. These results indicated that rainbow trout (a carnivorous

species) utilize lipids more effectively than carbohydrates as an energy source. In

contrast with rainbow trout, the omnivorous carp was able to utilize carbohydrates

effectively as an energy source. Excess carbohydrates in the diet of some carnivorous

fish have been associated with a decrease in health due to glycogen accumulation in the

liver (Gatlin III 2002). Therefore, it was expected that H. ahli (a carnivore) like rainbow

trout, would not have been able to tolerate high levels of carbohydrate. Histological tests

confirmed that there was an accumulation of glycogen granules in the vacuoles of the

hepatocytes of H. ahli on Diet 5 (54% carbohydrate), while P. socolofi had a few

granules on the periphery of the hepatocytes (Figure 4-5A). Percent LDA of both species

on Diet 5 was higher than the other diets. Normal hepatocytes have no observable

glycogen vacuoles, but the higher the glycogen content the greater the hepatocyte

vacuolation (Post 1987). Though the high mortalities observed in salmon and trout diets

on high carbohydrate diets (Phillips et al. 1948) were not observed in either P. socolofi or

H. ahli on Diet 5, growth was reduced and abdominal bloating and increased respiration

was observed in H. ahli on this diet (symptoms similar to that of Malawi Bloat [Ferguson

1989; Dixon et al. 1997; Smith 2000]). All the other diets did not produce observable

glycogen particles in the vacuoles. This suggests that for this study both species seem to

be able to tolerate up to 42% carbohydrate in the diet without adverse effects. These

results are similar to that observed by Chou and Shiau (1996) where hybrid tilapia

(Oreochromis niloticus X Oreochromis aureus) were able to perform well on diets

containing no greater than 42% corn starch.









The conclusion that juvenile African cichlids do not require high lipid levels as an

energy source for growth indicates that they can efficiently utilize digestible

carbohydrates as an energy source. This ability to utilize dietary carbohydrates for

energy is a major benefit for formulating diets. Increasing dietary carbohydrates, and

reducing the amount of lipid in the diet, in turn will decrease the amount of lipid storage

in the liver. Carbohydrates are considered the least expensive form of dietary energy

(Hertrampf and Piedad-Pascual 2000). In addition, carbohydrate is a source of metabolic

intermediates for synthesis of non-essential amino acids, and can have the added benefit

of improving the pelleting quality of steam-pelleted and extruded feeds (Robinson and

Wilson 1985). Therefore, some form of digestible carbohydrate can spare protein and

based on the diets tested can be included at levels between 18 to 42% in the diet. Further

research in carbohydrate utilization is warranted with African cichlids.

Feed Conversion Ratio (FCR) is a parameter that is used by farmers in the food

fish industry to estimate the amount of feed used to produce one kg of fish (De Silva and

Anderson 1995). Due to the small sizes of the juvenile African cichlids used in this

study, the FCR was high (> 2) compared to what is observed for farm raised food fish

(De Silva and Anderson 1995). Within the African cichlid industry, there is less concern

for feed conversion in the production of juveniles compared to the food fish industry. As

a result of the high selling price of these fishes (Chapter 1), a lower FCR resulting in

reduced feed costs may not have much effect on profitability.

In summary, both species had similar requirements for dietary protein, lipid and

energy. As hypothesized, the carnivorous H. ahli tolerated higher levels of dietary lipid

than the omnivorous P. socolofi, and P. socolofi tolerated higher levels of dietary








carbohydrates than H. ahli. Sexual maturity of P. socolofi and H. ahli was enhanced, and

growth and FCR was best for P. socolofi, on Diet 1. Although small juvenile P. socolofi

can be fed a high protein and lipid diet for the first 8 to 12 weeks of grow out (Chapter 2),

feeding this type of diet right up to the maturation stage (as shown in this trial) produced

fatty fish with fatty livers. It is suspected that under the practical conditions of

commercial production of African cichlids (as opposed to sterile experimental

conditions), excess fat in the liver and body cavity could affect the health and fecundity

of these fish. Reducing the lipid content from 19% (Diet 1) to 9.5 % (Diet 6) resulted in

a significant reduction in the lipid content of the body and liver of both species. Since the

high-lipid diets for African cichlids may compromise the health of these fish, under

practical conditions a diet containing less than 10% lipid is recommended for long term

feeding. Lowering the protein content to -36% (Chapter 2) can also be an added benefit

in improving water quality, and may also reduce feed costs without having an effect on

the target length and weight needed to reach marketable size for these fish. In

conclusion, the minimum dietary protein, lipid and energy levels for the growth of

juvenile African cichlids, without pathological changes in the liver, was found to be 36%

CP, 9.5% lipid and 4000 kcal/g, respectively.











-Table 4-1. Formulation and composition (g / 100 g dry weight) of experimental diets
Diets
Ingredients 1 2 3 4 5 6
Fishmeala 25.0 25.0 25.0 25.0 25.0 25.0
Casein 35.0 35.0 35.0 14.0 14.0 14.0
Wheat starch 20.0 20.0 20.0 40.0 40.0 40.0
Fish oilb 6.5 4.0 1.5 6.5 4.0 1.5
Soybean oilb 6.5 4.0 1.5 6.5 4.0 1.5
Stay Cc 0.4 0.4 0.4 0.4 0.4 0.4
Choline chloride 0.8 0.8 0.8 0.8 0.8 0.8
Vitamin premixd 1.0 1.0 1.0 1.0 1.0 1.0
Mineral premixe 1.0 1.0 1.0 1.0 1.0 1.0
Carboxymethylcellulose 3.0 3.0 3.0 3.0 3.0 3.0
Cellulose 0.8 5.8 10.8 1.8 6.8 11.8
Proximate Analysis
Moisture 5.53 5.53 5.20 6.50 6.50 6.50
Crude protein 55.09 56.58 56.24 35.95 35.38 35.59
Lipid 19.4 13.03 8.32 20.1 10.98 9.53
Carbohydratef 18.82 19.65 21.31 34.54 53.97 42.23
Ash 5.77 5.77 5.12 5.60 4.57 7.50
Crude fiber 0.92 4.97 9.01 3.81 3.04 5.15
Gross energy (kcal/kg) 5300 5100 4266 4990 4255 4000
Gross energy to protein
ratio 9.6 9.0 7.6 13.9 12.0 11.2
a Menhaden (68% crude protein), Omega Protein, Hammond, LA
b Fish oil and soybean oil, Sigma-Aldrich Co
c Stay C-35 (ascorbyl-2-polyphospahte) Hoffman-La Roche, Ltd.
d Federal Vitamin #30 by Zeigler Bros. Inc, Gardners, PA. To supply per kg of diet:
D-Calcium Pantothenate, 106 mg; Pyridoxine, 30.8 mg; Riboflavin, 52.8 mg;
Niacinamide, 220.4 mg; Folic Acid, 848 mg; Thiamin, 32.5 mg; Biotin, 0.35 mg; Vit B12,
0.22 mg; Sodium Menadione Biosulfate (Vit K), 11.4 mg; Vit E, 352.8 I.U.; Vit. D3, 441
I.U.; Vit. A, 6,615 I.U.
e Mineral premix #3 by Zeigler Bros. Inc. Gardners, PA. To supply per kg of diet:
ZnSO4, 165 mg; MnSO4, 27.5 mg; CuSO4, 5.5 mg; FeSO4, 48.4 mg
All other ingredients from ICN Biomedical Inc
fCarbohydrate = 100 (moisture+protein+lipid+fiber+ash)







68






f ;
c cSo0 .N



S0 00 0 Cl 0



"t 00



m 0; 0 ON













to
| 6 C S a 0 e N







c 0 00 :t Cl Q












oo ci
?>i t \" i





Sl & 0 3
00 0 8 5











4 >


+UO r 4j


- *I







ro El^

- ? .3 .* 5 U U o3

H3 S -








69





CO OC-4 N-t


o C) .. ...
mN N
.3 Cd M -0 C* ; m d m





N0 000 ,-0 N
S 0- m %r- ri V ) ri -n






c ^ -c
0 O O 0 OW 0 0







-, ^0 0 0 0 6 0 6 !
3 0- '00 .0 06


N N 0 C t(o -0
c)O
C: c c 0 O


C4-

0 0 0 '- '-
Cd Cm z o0 C-4 V







n kO O7 0 M N
C) 00
06 k- ,6







no ot


-Z 0 0 0 ",0 0 0 0 0
N-'j ('-'~ 0 () N







a <- .. .. 9 A








7d .
0) -
o a ca .0




O C C ON 0 M '


0 0 00 N 0O I0 N
(3 '-' z r^oo o




0 ONC- 0 0 0 O 0 >
:r t4 C i o0 J. -r \.
^ N t fN ^ t r o m C


6 1 iooos,666 ^

ri S ri Et t r CS
o^ -4 Cv i mv 0 ii N N3 0.)






~~~~~- vOt0 .. C

-z9
0) a
~~~ 0) 0) 0 0 -IC
H E




















C.

a.)
C.


Cm

0,
C)

c Y


00












a)




e-
C















0
a)










C.
0
co
so



a
-o
0


o




a
a
ci



c














o '
I r


c^ .
&-o\
000,
d..6






cM"i


to














a.a
e a
n I-



























t o
E o






a r


















4 A
I- l
X -



C)



o A
I O
o a .

I-,









0
e ,

f0c




3 <














- 5
1-u u. u 1

zu g < O t





Sirl>~l
c % 5S<


























a0 o
oo
00 C4 m Cr l ( 4 m# Cl







Cd
"2 o o
Aa
-db cq
1 CI

I.#tl N-- -

0 W )N V )0
^ o a o





0- oo o o
S.0
ff l 0- d d a \0 0 -
o .o! c oo S o
o--m C t oo M C
-- 00 Cn II





o on 0c
o o cc o c -o




..o V)0 ) 0 (
S; x^ d ooooSa -







'--H


o .

( r m 0 : 4) P v 0
.C N E/ o" '0 N C 4)
V)uo S Ed Ln w d c

0 w a1 N C ) 5- 4) a) -ij

-O 00 l O 4) o


40 Os C) N c 0I ( ~ I -
t n k m mn C d


1C C I a) 00 (9) C S 2 Q 2 2



r- ) m -n I1t 'o N 0 n













- 55% CP; 19% lipid
- .l- 57% CP; 13% lipid
.-.'.!.+-56% CP; 8% lipid
- -W 36% CP; 20% lipid
- 35% CP; 11% lipid
-- 36% CP; 10% lipid


-A


2 4 6


Time (weeks)


Mean growth of P. socolofi fed experimental diets with varying dietary
protein (%) and lipid levels (%) for 8 weeks


4.00

3.50

3.00
0)
2.50

( 2.00

1.50

1.00

0.50

0.00


Figure 4-1.











-- 55% CP; 19% lipid
- 57% CP; 13% lipid
- f+ 56% CP; 8% lipid
- -M 36% CP; 20% lipid
- -- 35% CP; 11% lipid
-- --36%CP; 10% lipid


~ -
-" -B .t?


0.5

0.0


4
Time (weeks)


Figure 4-2. Mean growth of H. ahli fed experimental diets with different with varying
dietary protein (%) and lipid levels (%) for 8 weeks


3.0

I 2.5
o

S2.0


1.5

1.0













































Figure 4-3. H&E stained liver sections of P. socolofi fed experimental diets for 8 weeks:
(A) Diet 1, representative of liver pathology code '3'. Hepatocytes are distended with fat
and nuclei are indistinguishable; (B) Diet 2; (C) Diet 3; (D) Diet 4. Note small,
hypochromatic nuclei at periphery of cell (arrows); (E) Diet 5; and (F) Diet 6. L= lipid
vacuoles, G= glycogen vacuoles. Note arrows showing nuclei at periphery of
hepatocytes.









































Figure 4-4.. H&E stained liver sections from the livers of H.
ahli fed experimental diets for 8 weeks: (A) Diet 1,
representative of liver pathology code '2'; (B) Diet 2 showing
severe vacuolation and representative of liver pathology code
'3'with hepatocytes distended with fat; (C) Diet 3, note that cells
are more defined; (D) Diet 4; (E) Diet 5, shows polygonal shaped
hepatocytes; and (F) Diet 6. L= lipid vacuoles. Note arrows
showing nuclei at periphery of hepatocytes.

















D


t-9
&si ^J


Figure 4-5. PAS stained liver sections ofP. socolofi and H. ahli fed
experimental diets for 8 weeks: (A) P. socolofi on Diet 5; (B) H. ahli on
Diet 5, showing glycogen accumulation in the vacuoles; (C) P. socolofi
on Diet 6; and (D) H. ahli on Diet 6. V = vacuoles, G = glycogen
granules. Note arrows showing nuclei at periphery ofhepatocytes.














CHAPTER 5
FORMULATION OF A MODEL DIET FOR THE COMMERCIAL PRODUCTION OF
JUVENILE AFRICAN CICHLID FISHES


Introduction

The success of the ornamental African cichlid industry has relied on the fact that

larvae are able to consume artificial diets at first feeding. Many begin feeding before the

yolk sac has been depleted and by 2 weeks post-hatch, juveniles are able to feed on a

200 starter diet. Although the use of artificial feeds to raise juvenile African cichlids is

extensive, presently no feeds are in use that are made specifically for their commercial

production. With the increase in intensive production, interest has arisen from producers

in finding a diet specific for African cichlids to improve their health, efficient nutrient

utilization and water quality. As reported in the previous chapters, feeding diets used by

the salmon and trout industry to cichlids led to rapid growth rates, but the resulting

health, protein utilization and nutrient waste reduction is far from optional.

Currently feed contributes to approximately 30 to 40% (Florida Tropical Fish

Farms Association [FTFFA] store, Gibsonton, Florida) of the total operational costs in

producing African cichlids; and two of the commercial feeds used by the farms are

Aquamax 400TM (50% CP, 15% fat) Purina Mills Inc., St. Louis, Missouri which costs

about $0.48 per pound and Zeigler trout starterTM (40% CP, 12% fat) Zeigler Brothers

Inc., Gardeners Pennsylvania, which costs $0.47 per pound (FTFFA). Unlike the food

fish industry where fish is sold by the pound, ornamental fish are sold by the number of









individual fish (otherwise known as pieces) being purchased and priced according to the

length or life stage of the fish (e.g. brood stock). Based on the present price per fish (at

least $0.75 for a 5 cm [2 inches] fish, FTFFA), the value of ornamental African cichlids

is many times higher than the cost of feed. For example, data from Chapter 4 shows that

an average of 5.0 g of feed would be required to raise 12 juvenile Pseudotropehus

socolofi from 4.0 cm (1.5") to 6.0 cm (2.5") for 8 weeks. Therefore 4.16 kg (9.16 lb) of

feed would be required to produce 10,000 fish for 8 weeks. These costs may vary

according to the farming practices. At a cost of $0.48 per pound of feed, for 8 weeks, the

farmer would spend $4.39 in feed costs to grow-out 10,000 fish. Although the feed costs

may seem relatively small, under commercial production, these feed costs will have an

effect on profitability. The challenge is to produce a specialty diet (one that will decrease

fat accumulation in fish while still providing for growth rates similar to the feeds being

used presently) at a cost less than or similar to what is being fed presently.

Nutrient Requirements and Feed Ingredients

The increased understanding of the nutritional requirements for various fish

species and technological advances in feed manufacturing has allowed the development

and use of manufactured or artificial diets (formulated feeds) to supplement or to replace

natural foods in the aquaculture industry. Formulation of a practical diet for intensive

fish culture is based upon the knowledge of the nutritional requirement of the fish species

and life stage, in particular the optimum protein and energy levels. Nutrients essential to

fish are the same as those required by most other animals: water, proteins (amino acids),

lipids (fats, oils, fatty acids), carbohydrates (sugars, starch), minerals, and vitamins. In

addition, pigments (carotenoids) are commonly added to the diet of ornamental aquarium









fishes to enhance coloration. Based on the previous experiments (Chapters 2 to 4) for

African cichlids and what is known for fish in general (New 1987; Halver 1989; NRC

1993; De Silva and Anderson 1995; Lovell 2002), the proportions of various nutrients

that would be optimal in a standard diet for juvenile African cichlids are given in Table 5-

1. Proteins (which are comprised of amino acids) are the major organic material in fish

tissue and make up 65 to 75% of the total tissue on a dry-weight basis (Wilson 2002).

Fish require a balanced mixture of both essential and nonessential amino acids for tissue

growth, maintenance and reproduction (NRC 1993). Lipids refer to substances known as

fats and oils and are a major source of energy for fish. In addition, lipids have other

important functions such as providing structure for cell walls, serving as a carrier for the

fat soluble vitamins A, D, E and K, supply essential fatty acids and are precursors of

hormones (NRC 1993; Hardy and Barrows 2002). Dietary carbohydrates, although not

required by fish, can provide a cheap source of energy to those fish (such as African

cichlids) that are able to utilize carbohydrates.

A wide range of raw ingredients are available for diet formulation for aquatic

species. Nutrient analyses and recommended usage rates of these ingredients are

provided in many references (New 1987; Winfree 1992; New et al. 1993; NRC 1993;

Hertramph and Piedad-Pascual 2000). Ingredients used in fish feeds such as fishmeal,

soybean meal, corn gluten meal, brewer's yeast, and wheat gluten are excellent sources of

crude protein. Oils from marine fish such as menhaden, and vegetable oils from soybean,

canola, sunflower, and linseed are common sources of lipids. Menhaden oil and linseed

oil are excellent sources of omega-3 fatty acids while soybean and canola provide omega-

6 fatty acids. Along with oils, other inexpensive sources of energy that may spare protein








from being used as an energy source are cooked carbohydrates from grain by-products.

Rice bran and the flour of corn, wheat or other 'breakfast' cereals are excellent examples

(NRC 1993; Hertramph and Piedad-Pascual 2000).

Leafy green vegetables are a typical form of natural foods used in ornamental fish

diets. Although vegetables are composed mainly of water, they contain minerals,

carbohydrates, and vitamins. Kale, zucchini, and spinach, are some examples of

nutritious vegetables. Vitamins and minerals are usually prepared synthetically and

available commercially as a balanced and pre-measured mixture known as a vitamin or

mineral premix. This premix is added to the diet in small amounts to furnish the levels of

vitamins and minerals required by fish (NRC 1993).

Fiber and ash (minerals) are a group of materials found in diets and diet

ingredients (NRC 1993). Fiber is the indigestible plant matter such as cellulose,

hemicellulose, lignin, pentosans and other complex carbohydrates found in feedstuffs

(NRC 1993). The digestibility of crude fiber by most fish species is less than 10%.

Higher concentrations will reduce feed intake, affect the utilization of nutrients and

depress growth (NRC 1993; Hardy and Barrows 2002). Although crude fiber levels

should not exceed 10% in fish feeds, fiber levels between 3 to 5% are preferred. Some

fiber is useful as it provides bulk and facilitates the passage of food through the intestinal

tract (NRC 1993; Hardy and Barrows 2002). Ash is a source of calcium and phosphorus

from bone or plant material and in practical diets both should be no greater than 8 to 12%

of the formulation. A high ash content also reduces digestibility of ingredients in the diet

resulting in poorer fish growth.








Carotenoids, a group of naturally occurring lipid soluble pigments, are responsible

for the variations of yellow, orange and red coloration found in cichlids. When attached

to proteins or lipids, carotenoids result in other colors such as blue, green, purple or

brown. These pigments are produced primarily by plants and algae and must be

supplemented in the diet of fish as they are unable to synthesize carotenoids (Guillaume

2001). The addition of pigments to the diet of African cichlids is important due to the

vast array of colors shown by these fish. Aside from the salmon industry, the commercial

feeds used by the farmers usually do not contain a suitable concentration and variety of

carotenoids needed for the enhancement of coloration. If carotenoids are not constantly

supplemented, coloration will diminish over time reducing the marketability of the

cichlids.

A variety of natural and synthetic pigments are available for inclusion into fish

diets. Astaxanthin is the most frequently occurring carotenoid found in aquatic animals

and is the most commonly used dietary pigment. Natural sources of astaxanthin can be

found in various plant materials such as algal meals (blue-green algae Spirulina spp.,

kelp); single cell proteins such as Phaffia yeast; marine animal meals (fish, shrimp, and

crab); fish oil such as salmon oil and other aquatic animals (krill, freeze-dried brine

shrimp and copepods). Synthetic astaxanthin (obtained from companies such as

CyanotechO Corporation, Kailua-Kona, Hawaii and BASF Corporation, Mount Olive,

New Jersey) can be added at 50 to 150 mg/kg to enhance pigmentation in ornamental

fishes (Hertramf and Piedad-Pascual 2000). Carotenoids can also be found in extracts

from plants such as French marigold petals (Tagetes sp.) and red peppers (paprika).

Venders (e.g. Alcosa International Inc., San Antonio, Texas) for marigold petals suggest









that adding the pigment at 2.5 to 5% of the diet aids in the yellow coloration offish. The

carotenoids lutein and zeaxanthin, which are found in corn and corn by-products such as

gluten meals, also impart a yellow color to fish (Guillaume 2001; Hertramph and Piedad-

Pascual 2000). The metabolism and deposition of these carotenoids varies from species

to species.

In the literature, there are no studies that investigate the effects of various pigment

sources specifically on ornamental African cichlid coloration. However, the use of

Spirulina algae in the diet as a source of pigment is common in the ornamental fish

industry (personal communication, FTFFA). Ako and Tamaru (1999) found that adding

2% Spirulina to the diet of a South American cichlid was effective in enhancing

coloration. Based on recommendations by Lovell 1992, adding 500 mg of astaxanthin

and/or Spirulina mixed with 30g of fish oil to 2.2 kg of Nelson's SilvercupTM, (Nelson

and Sons Inc., Murray, Utah) fed at 5% of body weight twice per day, enhanced

coloration in adult red empress cichlids (Aulonocara nyassae) after 2 weeks (personal

observation). Boonyaratpalin and Unprasert (1989) observed an increase in the

coloration in fins and tail oftilapia (Orechromis niloticus) after 2 weeks when fed

Spirulina at 10% of the basal diet or marigold petals at 5% of the basal diet. When

shrimp head meal was fed at 15% of the basal diet, improved coloration was observed

after 4 weeks. A deep yellow color was produced in Gold tilapia (Orechromis

mossambeca) when fed a diet containing 50 ppm marigold petals for 8 weeks while,

capsicum extract fed at 200 ppm produced a light orange color (Lovell 1992). It is

important to note that the higher the level of inherent carotenoids in ingredients such as,

corn gluten meal, krill or shrimp meal, the less supplemental carotenoids are needed.








Another important ingredient in fish diets is a binder to provide stability to the

pellet and reduce leaching of nutrients into the water. Beef heart has traditionally been

used both as a source of protein and as an effective binder in farm-made feeds (Appendix

A). Carbohydrates (starch, cellulose, pectin) and various other polysaccharides, such as

extracts or derivatives from animals (gelatin), plants (gum arabic, locust bean), and

seaweeds (agar, carageenin, and other alginates), are also popular binding agents (Meyers

and Aein-Eldin 1972).

Preservatives, such as antimicrobials and antioxidants, are often added to extend

the shelf life of fish diets and reduce the rancidity of lipids in the diet. In Florida, these

preservatives are of special importance due to high temperatures (above 280C) and

humidity (80 to 100%) encountered in most storage depots. Moisture enhances the

growth of toxic molds and high temperatures contribute to rancidity of oils and the

deterioration of vitamins, amino acids and pigments in feeds. For a review of nutritional

consequences and pathological effects of poor storage, see Chow (1980), New (1987),

and Barrows and Hardy (2001). Vitamin E is an effective, but expensive, antioxidant

(Hardy and Barrows 2002) that can be used in farm-made formulations. Commonly

available commercial antioxidants are butylated hydroxyanisole (BHA), butylated

hydroxytoluene (BHT), and ethoxyquin (Hardy and Barrows 2002). Butylated

hydroxyanisole and BHT are added at 0.005% of the dry weight of the diet or no more

than 0.02% of the fat content in the diet. Depending on the total ethoxyquin level in the

diet, ethoxyquin is added at a rate of 0.0125% of the diet in order to achieve a maximum

level of 150 ppm. Sodium and calcium salts ofpropionic, benzoic or sorbic acids are








commonly available antimicrobials added at a level of 0.025% of the diet (Robinson and

Wilson 1985; Barrows and Hardy 2001).

Other common additives incorporated into feeds are chemoattractants and

flavorings. Natural feed ingredients (liver, fish, shrimp, and krill meal), by-products

from natural foods (condensed fish solubles added at 5% of the diet and fish

hydrosylates), synthetic amino acids (L-glycine, L-alanine, L-glutamic acid) and the

"vitamin-like" nutrient betaine are also known to stimulate strong feeding behavior in

fish. All are incorporated into fish diets to enhance palatability and feed intake (Carr et

al. 1996; Barrows and Hardy 2001).

Feeds can also be formulated to enhance the nonspecific immune responses by

supplementing with ingredients such as beta glucans, chitin, dried Spirulina algae and

brewer' yeast (Saccharomyces cerevisiae). These ingredients have been found to

increase the phagocytic activities ofmacrophages and neutrophils in channel catfish

(Gatlin III 2002). Brewer's yeast is the foremost natural source of B vitamins and has a

high content of digestible protein. Although it is expensive, since the 1970's Spirulina

has been used in the industry to promote coloration in aquarium fish. It is a

commercially produced microalgae and is rich in amino acids, pigments, and vitamins.

The typical nutrient profile of Spirulina can be found in Henson (1990) and has the

following benefits: better growth rates, enhanced palatability of feeds, improved quality,

coloration, better survival rates and reduced medication requirements. Spirulina was

reported to also improve the beneficial intestinal flora, which enhances resistance to

infection by displacing harmful bacteria Henson (1990). These immunostimulants are









not restricted by the U.S. Food and Drug Administration (FDA) regulations and

suggested levels for use range from 0.2% to 2.7% of the diet.

Feed Formulation and Preparation of a Juvenile African Cichlid Diet

With a few exceptions (e.g. unaltered live foods), feeding a single type of

feedstuff is neither complete nor balanced and does not supply all the nutrients a fish

needs in its diet (New 1987). Hence, two or more ingredients should be combined to

make homemade, laboratory, or commercial feeds. Formulated feeds are a mixture of

ingredients that are processed into pellets, granules, flakes or other forms. They are

formulated to meet all of the nutritional needs of fish in tank or race-way culture systems,

since natural food is limiting or absent in these systems. These feeds are known as

complete feeds. For fish in pond systems, only a portion of the nutritional needs is met

by supplemental feeds (Barrows and Hardy 2001). Complete feeds are formulated to

have a final moisture content depending on the type of feed desired ranging from: dry (6

to 10% water content), semi-moist (35 to 40% water) or wet (50 to 70% water) (Goddard

1996). Most feeds used in intensive production systems or in home aquaria are

commercially produced as dry feeds. Dry feeds contain less water and more lipid,

protein, vitamins and minerals than do natural foods (De Silva and Anderson 1995). Dry

feeds may consist of simple loose mixtures of dry ingredients such as, 'mash or meals',

or more complex compressed pellets or granules. Often, pellets are broken into smaller

sizes known as crumbles. The pellets or granules can be made to sink or float by cooking

with steam or by extrusion. Although floating or sinking flakes are another form of dry

foods used as popular diet for aquarium fishes, crumbles or pellets are the more

recognized form of feed for commercial aquaculture. Their use results in improved








bioavailabilty of the nutrients in the raw ingredients, higher feed intake, feed efficiency,

less waste and easier storage (Hardy and Barrows 2002). There are many texts available

on the commercial manufacture of pelleted aquatic feeds (New 1987; New et al. 1993;

Goddard 1996; Chang and Wang 1999; Jobling et al. 2001; Hardy and Barrows 2002).

Feed Characteristics

To obtain the best results in feeding, not only should the diets be nutritionally

complete, but they must also be highly digestible, palatable, be readily consumed and

ingested in a sufficient quantity for growth and survival. Fish feeds should also have

good water stability (remain intact for at least 15 to 30 minutes without leaching of

nutrients), must be of the proper pellet size (25-50% of the mouth diameter) (Tucker

1998) and texture, and not cause pathological change in the animal (Wantanabe 1998).

Feed acceptance depends on its availability, appearance, particle size, smell, taste and

texture (FAO 1980; Jobling et al. 2001). The ability to detect and ingest feed is affected

by pellet bulk density (sinking rate), size (shape, diameter and length), color and texture

(hardness). If the diet is unacceptable, growth rates, health and quality of fish are

affected. In addition, excessive waste due to uneaten feed and increased fecal material

resulting in an increase in ammonia production can cause a breakdown in water quality,

which increases the susceptibility of successive disease outbreaks.

Feed Formulation Methods

In general, diets are formulated based on a least-cost analysis (i.e. reformulating

the diet when feed ingredient prices change) and are formulated manually or by the use of

a computer program (Hardy 1989). On an individual farm, basic formulations can be

done manually with the Pearson square formulation (FAO 1980; New 1987). The








"square" method is helpful for quick, on-the farm formulations and works best if only a

few ingredients are used. In many fish diets, protein is the most expensive portion and is

usually the first nutrient that is computed in the diet formulation (Hardy 1989). This

method can easily determine the proper dietary proportions of high and low protein

feedstuffs to add to the diet in order to meet the dietary requirement of the fish being fed.

Hardy (1989) provides examples of using rice bran and soybean meal to formulate a diet

containing 25% crude protein. The square method can also be used to calculate the

proportion of feedstuffs to be mixed together to achieve a desired energy level as well as

a crude protein level. However, it cannot be used to simultaneously solve for both

protein and energy.

Computer software such as Brill Feed Management Systems (Brooklyn Center,

Minnesota) or Dalex Computer Systems Inc. (Waconia, Minnesota) are used in the feed

industry to calculate least-cost formulations by linear programming and involves the

simultaneous solution of a series of linear equations (Hardy and Barrows 2002). These

programs are expensive, and are only economically feasible when very large numbers

and volumes of ingredients are being used regularly and the farm or feed mill has the

capacity to purchase and/or store these ingredients (Robinson and Wilson 1985; New

1987). A Microsoft Office Excel (Microsoft Corporation, Redmond, Washington)

spreadsheet can also be used to formulate simple diets and is useful in estimating the cost

and amount to mix for each ingredient. Unlike formulating with a software package, the

nutrient profile is not as extensive and only takes into account a few nutrients. Therefore

the calculated output may not be as accurate (Table 5-3) as that found with computer

formulations (Table 5-4). The following information is required for formulation: (1) a









list of available ingredients, their nutrient composition and cost; (2) knowledge of the

quality, suitability and limits of the ingredients to the fish being fed; and (3) the nutrient

requirements of the fish being fed (New 1987; Hardy and Barrows 2002). Based on the

results of the growth performance and health of juvenile African cichlids (Chapters 3 and

4), a diet can be formulated for juvenile African cichlids that contains between 36% to

40% CP, and 7 tol0 % fat (Table 5-5). Table 5-2 describes the nutrient composition for

each ingredient to be used to formulate this diet and their costs. Using Table 5-2 and the

knowledge of the nutrient requirements of the fish, a diet can now be formulated for the

commercial manufacture of an extruded pelleted feed using both an Excel spreadsheet

program (Table 5-3) and also with the Dalex Computor Systems software (Table 5-4).

Diet formulations usually have restrictions, which are classified as nutritional,

processing, inherent problems with feedstuffs and miscellaneous (Robinson and Wilson

1985). For example, it has been found that more than 20% soybean meal in the diet can

be toxic to fish due to antinutritional factors such as trypsin concentrations. An example

of restrictions for feed processing is the inclusion of adequate amounts of starch for

sufficient binding of ingredients when making pellets. The specifications for juvenile

African cichlid diets are presented in Table 5-5. A maximum, minimum and target

percent inclusion limit is suggested for each nutrient formulated and for specific feed

ingredients. These specifications are determined from the previous experiments and for

those of channel catfish and tilapia (Winfree and Stickney 1981; Robinson and Wilson

1985; New 1987; Pesti and Miller 1993; Hertramph and Piedad-Pascual 2000) and can be

used as a guide for formulation of diets for African cichlids.








In Table 5-3, column 2 is the percentage of each ingredient used in the mixing of

one pound of the diet. Columns 3 to 6 are the percentages of each nutrient in the

ingredient contributing to the diet and column 7 is the cost per lb of each ingredient in the

diet. For example, from Tables 5-2 and 5-3, an inclusion rate of 38.7% fish meal costing

$0.39 per Ib, would contribute 0.387*0.39 = $0.15 per lb to the cost of each pound of diet

formulated. Whereas, an inclusion rate of 38.65% fish meal, contributes 23.96% (62%

crude protein in fish meal 0.3865) crude protein to the diet. These results are then

totaled to see how they comply with the specifications of the ration. In Table 5-4 the

same ingredients have been used to formulate the diet using the Dalex program. This

program is able to provide more accurate information about the cost and nutrient analysis

of the diet compared to the Excel program. For example, using the computer program,

the cost for formulating an African cichlid diet based on the ingredients used in Table 5-2

is $0.23 per pound (Table 5-4) versus $0.33 per pound using the Excel program (Table 5-

3). Tables 5-3 and 5-4 are examples of one version of a diet that can be formulated for

juvenile African cichlids. With the spreadsheet or computer program, re-adjustments can

be made easily and ingredients can be substituted for others. For a list of ingredients that

can be used to substitute for the ones used in Tables 5-3 and 5-4, refer to NRC (1993) and

Hertramph and Piedad-Pascual (2000).

Considerations for the Preparation of Farm-Made Feeds versus the Use of
Commercial Feeds by Farmers
With information on how to formulate a diet, the farmer is able to decide whether

it is economical to prepare his own feed or to purchase commercial feeds. Most farmers

find it more convenient to use commercial feeds because of the lack of capital to invest in

equipment and or the unavailability of the feed ingredients. Although feed costs are 30 to








40% of rearing African cichlids, there are many advantages to purchasing ready-made

feeds. The feed quality of commercial feeds can be controlled in terms of uniformity,

and nutrient composition. They are also drier and have a longer shelf life than farm-made

feeds. However, many commercial aquafeeds are over-formulated with nutrients and

generally tend to have excessive amounts of protein and energy creating the fatty liver

syndrome observed in African cichlids. Total phosphorous levels may also be high and

not available to the fish thereby increasing unwanted algae growth in the aquatic system.

Farm-made feeds are more economical for farms that have access to inexpensive

feed ingredients, can invest in the equipment needed, and have the time and/or labor to

prepare their own feed (Chong 1995; Piedad-Pascual 1994). The advantage of making

their own feed is the ability of the farmers to take more control over the type of

ingredients needed to meet specific requirements. For example, addition of carotenoids

before shipment, adding Vitamin premix to a vitamin deficient diet, or controlling the

amount of lipid or protein being added to the diet.

The equipment used for making feed on the farm depends on the size of the fish

and the quantity of feed that is desired. The primary use of equipment is for size

reduction of ingredients, blending, forming, and drying. Table 5-9 specifies the types of

equipment that can be used for each processing operation for the manufacture of farm-

made feeds (Wood 1993). The choice of equipment will depend on the scale of feed

manufactured. More information of the description and use of each piece of equipment

can be obtained from New (1987). The capital costs of equipment depends on the

volume of feed produced while the equipment cost per pound of feed produced will vary

based on the initial equipment price, maintenance costs, financing costs, and rate of