Ichthyology: The Science

The study of fishes, broadly termed "ichthyology," en Systematics and Taxonomy of Fishes
compasses many aspects. Some ichthyologists are most
concerned with studies of the diversity, distribution, and Systematics is the theory or practice of discerning the
interrelationships of fishes. Others concentrate on the orderliness of nature or other systems. Taxonomy is the
physiology or functional morphology of fishes, seeking theory and practice of classifying that orderliness in a
to determine how the various body parts of fishes inter worded system (Mayr, 1969). Systematists work with
act to facilitate feeding, locomotion, respiration, or oth units, collectively called "taxa," which may consist of
er vital functions. Studies of fish genetics have become the smallest recognizable units (usually a population or
numerous in recent years and offer much potential infor group of populations), often referred to as species, sub
mation to those studying both the diversity and relation species, or races, or they may consist of groupings of
ships of fishes as well as insights for those attempting to species into more inclusive taxa such as genera, fami
manage fish population resources. Still other ichthyolo lies, and orders. Information follows on nomenclature
gists specialize in behavioral studies, gathering and an of these taxa, and additional information is given in the
alyzing both qualitative and quantitative data, based on section "How to Use the Family and Species Ac
observations of fishes in natural environments and counts." Readers who want to obtain information be
aquaria, to learn more of their movement patterns and yond what follows on systematics and taxonomy, as
behavior associated with feeding, courtship and spawn well as on the evolution of different ideas and ap
ing, schooling, protection, and other aspects. Studies proaches to systematics, and the controversies sur
on fish biology, including basic life histories and fish rounding them, should consult the works of Simpson
ecology (how fish interact with their environments), are (1961), Sokal and Sneath (1963), Hennig (1966),
very important to a better understanding of these ani Blackwelder (1967), Mayr (1969), Eldredge and
mals. Within the past couple of decades, a great deal of Cracraft (1980), Nelson and Platnick (1981), Wiley
emphasis has been placed on the study of the early life (1981), Lundberg and McDade (1990), and many pa
history (larval development) of fishes. Other kinds of pers published in the journals Systematic Zoology and
studies include those of fish population dynamics, man Cladistics.
agement techniques for economically important species,
propagation, fishing techniques, and fish diseases. Alpha Taxonomy and Zoological Nomenclature. In tax
Those who specialize in these latter, somewhat more onomy, the species is the most generally accepted basic
applied, aspects are most often referred to as "fisheries unit, and studies that strive to discern these basic units
biologists." Nearly all aspects of ichthyology or fishery are known as "alpha taxonomy. " Readers should be
biology are germane to the subject matter of this book aware that species concepts vary among taxonomists.
at some level, but those of central interest are the sys Also, some choose not to name units below the level of
tematics, taxonomy, and distribution of fishes, as well species while others subscribe to subspecies concepts,
as basic biological information. The following sections which may also differ among workers. Perhaps one of
are intended to acquaint the readers with these aspects the oldest and most widely adhered to concepts is that
in particular. Those wishing to find more extensive in of the "biological species" which defines a species as a
forn1ation on ichthyology or fishery biology should con group of interbreeding or potentially interbreeding pop
sult the works of Lagler (1956), Lagler et a1. (1977), ulations reproductively isolated from all others (Mayr,
Bond (1979), Moyle and Cech (1988), Nielsen and 1969). Such a concept is, of course, very difficult to test
10hnson (1983), and Schreck and Moyle (1990). Exten or perceive in nature for many groups of organisms,
sive bibliographic information on fishes can be found in such as freshwater fishes isolated in different drainage
Dean (1916-1923) and in the Pisces section of the Zoo basins. Subjective decisions must be made concerning
logical Record. whether some populations may be capable of inter-

ICHTHYOLOGY: THE SCIENCE 53

breeding, or "intergrading," with others, and subspecies the same locality as the holotype, either concurrently or
designations are often invoked under this concept. A at other times; topotypes have no formal taxonomic sta
second concept is that of "evolutionary species" (Simp tus but are occasionally employed in resolving nomen
son, 1961) which views species as representing lineages clatural problems in the absence of type material.
maintaining identities from other such lineages and hav In the system of binomial nomenclature formalized
ing their own evolutionary tendencies and historical by the Swedish biologist Carolus Linnaeus in the
fate. Polytypic (having subspecies) species are often l700s, the name of a species (always italicized or un
recognized under this concept as well. A third concept derlined) consists of the name (capitalized) of the genus
is that species are populations or groups of populations it is placed in, plus the species epithet (lowercase)
of monophyletic (common ancestral) origin which have (e. g. , Etheostoma blennioides). Subspecies names (tri
differentiated from others (sometimes referred to as the nomials) follow the species epithet (e. g. , Etheostoma
"phylogenetic" or "cladistic" concept) (Rosen, 1978, blennioides blennioides). When subspecies of a species
1979; Cracraft, 1983; Lundberg and McDade, 1990). are recognized and published by a taxonomist, the sub
Differentiation is detected by the presence of presumed species encompassing the population from which the
apomorphic (uniquely derived) traits. In practice, all of species was originally described must retain that spe
these concepts involve some degree of subjectivity; the cies's name, as in the above example, and is known as
species is thus an arbitrary unit of taxonomic conve the nominate subspecies, while other populations may
nience. No matter what the concept, in most cases what be given other names (e. g. , Etheostoma blennioides
taxonomists really must do is identify populations of or gutselli). Only names published since 1758, beginning
ganisms and what they feel to be consistent variation with Linnaeus's Systema Naturae, are considered valid
(morphological or genetic) between these populations, under the international rules of nomenclature (ICZN,
arbitrarily specify acceptable limits to variation in a 1985), which govern taxonomic practices. When a tax
species, and make decisions based on these limits, per onomist judges two previously published species to be
haps lending credence to Regan's (1926) statement that the same, the first (older) published name has priority
a species "is what a competent taxonomist says it is. " and the newer (more recent) name may be placed in
This, of course, has been a simplified explanation of synonymy in a subsequent revision.
species concepts; there are numerous, much more in
volved, philosophical treatments of such matters (e. g. , Classification. Systematists attempt to reflect evolution
Rosen, 1978; Wiley, 1978; Cracraft, 1983, 1987; ary relationships in hierarchical classification systems.
Loevtrup, 1987; Chandler and Gromko, 1989; de In the most basic classification system, modified from
Queiroz and Donoghue, 1988, 1990; several works in that first erected by Linnaeus in the 1700s, species hy
Ott and Endler, 1989; and Wheeler and Nixon, 1990). pothesized to be closely related are grouped within gen
A taxonomist who believes he or she has identified a era, related genera within families, families into orders,
species new to science gives it "official" status (accord and orders into classes. Beyond the basic classificatory
ing to the International Rules of Zoological Nomencla levels of the Linnaean hierarchy, the systematist may
ture) by publishing a formal description in the scientific choose to recognize intermediate-level catagories (e. g. ,
literature and giving it a unique name, the species subgenera, tribes, subfamilies, superfamilies, sub
epithet. In that publication, a museum specimen, called orders, or superorders). With respect to nomenclature,
a holotype, is designated to forever serve as the stan each genus has a type species to serve as the standard
dard of reference for the new species. Under older prac for that grouping and with which the generic name must
tices, multiple types (syntypes or type series) were always remain, regardless of what other species might
designated which, in some cases, later proved to com be assigned to it. A large percentage of the fish genera
prise more than one actual species, and the practice was ever described are listed in Jordan (1917-1920), Golvan
discontinued. Other specimens used in the description (1962), and Eschmeyer (1990). If subgenera are to be
may be designated as paratypes. When the type of a recognized, a nomenclatural convention similar to that
species is lost, or syntypes are polytypic, and clarifica of subspecies applies. The subgeneric grouping contain
tion is needed, a revising taxonomist may designate a ing the type species retains the name of that genus, as
new one, a neotype in the former case or a lectotype in for example, Etheostoma (Etheostoma) blennioides,
the latter, to serve the same function as the holotype. while other subgeneric groupings may receive other
Another term, topotype, is often seen in taxonomic lit names, as in Etheostoma (Nothonotus) rujilineatum;
erature and simply refers to specimens collected from species names are most often written as binomials, but

54 The Fishes of Tennessee

if the subgeneric allocation is to be indicated, it is cation, are deemed primitive or derived by comparison
placed in parentheses as above. In the literature, when a to an "outgroup," consisting of species presumed to be
genus and its nominate subgenus are being inter more distantly related than the organisms under study
changeably discussed, they may be distinguished by re are to one another. Conditions also found among the
ferring to the genus in its entirety as "sensu latu" (in the outgroup are considered to be primitive (i.e., attribut
broad sense) (e.g., Etheostoma s.l.) and the subgenus able to a distant common ancestor) while those that are
in the restricted sense, "sensu strictu" (Etheostoma not present in that group might be considered derived
s.s.). At higher levels, each family has a type genus. evolutionary novelties indicative of more recent com
Family names (all ending in "idae") are extensions of mon ancestry and, hence, closer relationship. Decisions
the type genus name (e.g., Perca-Percidae); names of must also be made, often aided by the sum of overall
orders stem from one of the component families and evidence in a phylogeny (branching scheme of relation
end in "iformes" (e.g., Perciformes). Like species, ge ships), as to whether similar attributes are homologous
nus and subgenus names and names of higher rank may (of common genetic origin) or homoplasous (of separate
be synonymized or resurrected as deemed necessary by genetic origin but similar because of evolutionary con
taxonomists publishing revisionary studies. vergence or happenstance); only the former can be con
There are three basic approaches or schools of sidered shared derived characters. Only derived
thought on how to arrive at classifications: evolution characters are used in support of hypotheses of close re
ary, phenetic, and phylogenetic. The relative merits of lationships and the identification of closest relatives or
these have been hotly contested among systematists in "sister groups"; those deemed to be primitive, while
recent years. Evolutionary systematists take a varied they may contribute to the overall similarity of the or
(sometimes referred to as "synthetic", e.g., Nelson, ganisms, are not considered important at the level of the
1984) approach to classification, placing taxa into ingroup. Though somewhat controversial (sec Nelson,
groups both according to relative similarity within and 1978, 1985; Mabee, 1989a, b), developmental (on
between groups, interpreted as the degree of change be togenetic) information is also sometimes used in charac
tween common ancestors and descendant groups, and, ter analysis, with those features that develop later in the
to the extent that these are based on a phylogenetic ontogeny of, for example, a larval fish, generally con
framework, on probable "derived" characters (see be sidered more derived. Whatever the method of charac
low). In some cases, these classifications may also be ter analysis, among the possible "phylogenetic trees"
influenced by opinions as to where taxa are most likely arrived at, the one with fewest branchings or steps (the
to fall into hypothetical lineages perceived in the fossil "most parsimonious") is usually accepted as the most
record or according to processes believed to be acting tenable hypothesis of relationships. Cladistic methodol
on these lineages, such as geological history. Pheneti ogy was introduced to the United States by Willi Hcn
cists construct groups purely on overall similarity nig (1965, 1966) and is gaining acceptance in recent
within and between groups with no regard to either his years among fish systematists as the preferred approach
torical information or the advanced or primitive natures to determining phylogenies on which to base truly natu
of those individual features which contribute to sim ral classifications.
ilarity. This method has been most commonly applied at Systematists may make two kinds of "mistakes" in
lower taxonomic levels (species relationships, etc.). On erecting classifications and naming groups. Rather than
the other hand, phylogenetic systematists, or cladists, attaining the desired monophyletic groupings, some
strive to construct groups defined strictly on what are classifications may result in polyphyletic or paraphyletic
believed to be shared evolutionary novelties, or "shared groupings. Polyphyletic (error of inclusion) groups con
derived" characters (synapomorphies), characters pas tain members which do not share most recent common
sed down through common lines of descent, which de ancestors, thus stemming from separate lines of descent
note the genealogy of the group. In this way they hope and not being one another's closest relatives. Paraphyle
to attain "natural" classifications based solely on mono tic (error of omission) groups do not contain all of the
phyletic groups, that is, groups descended from com descendants of a given ancestor necessary to result in a
mon ancestors. "complete" monophyletic group because some actual
In cladistic analyses, conditions, such as a particular descendants remain arbitrarily assigned to another
type of bony structure or color pattern, shared between group. Thus systematists must continually attempt to re
organisms under study which are candidates for inclu solve relationships among organisms and update classi
sion in a group (ingroup) at a particular level of classifi- fications to reflect these relationships in order to avoid

ICHTHYOLOGY: THE SCIENCE 55

 such groupings. Classifications are always provisional
and constantly changing to reflect new hypotheses. Sev
eral groupings, especially higher groupings of fishes,
are strictly provisional, erected for practical conve
nience, until such time as higher relationships are better
understood.
Classifications of fishes have evolved for over three
centuries, beginning even before the concepts published
mon ophylet c by Linnaeus (1758) in Systema Naturae, which were
based primarily on those of Linnaeus's fellow Swede,
9 r 0 U P Peter Artedi, and a few even earlier (mid- 1600s) con
cepts of Englishman Francis Willoughby. In the late
1700s and early 1800s, European ichthyologists greatly
expanded on the knowledge of Linnaeus's time and
published several comprehensive faunal works reflect
ing their ideas of classification. From 1798-1803, the
E French ichthyologist B. G. E . Lacepede published his
Histoire Naturelle des Poissons, and German ichthyolo
gists Marcus E. Bloch and Johann G. Schneider pub
lished Systema Ichthyologiae in 180 1. Later came one
of the great classifiers of the animal kingdom, the
Frenchman Georges Cuvier, who published Le Regne
Animal ( 18 16) and began the classic 22-volume work
paraphylet c
Histoire Naturelle des Poissons with Achille Valen
9 r 0 U P ciennes (1828- 1849). In the 1830s, the Swiss zoologist
Louis Agassiz (later of Harvard) published fish classifi
cations with particular emphasis on fossil forms, and
English ichthyogist William Swainson ( 1839) published
a work devoted strictly to fish classification; he may
have been the first to formalize the "idae" endings to
family names within fishes. Classifications of higher
groups of fishes were hypothesized by Mueller (1844)
and Agassiz ( 1862). More comprehensive works reflect
ing then-current views of classification ensued in the
following decades (Gill, 1872; Guenther, 1859-1870;
Boulenger, 1904). Jordan ( 1923) and Berg ( 1940) pub
lished major classifications based in a large part on pre
vious concepts of C. T. Regan (1909) of the British
polyphylet c
Museum. Jordan's classification included genera in ad
9 r 0 U P dition to families, an effort repeated by Norman ( 1957).
A new, considerably revised, provisional classification
was attempted by Greenwood et al. (1966) with a par
Diagrams (i.e., c1adograms) of actual evolutionary (phyloge
ticular focus on attempting to reflect the genealogy of
netic) relationships among five taxa, A-E, showing different
kinds of groupings a systematist may hypothesize: monophyl groups. Most recently, classifications somewhat mod
etic group-a correct hypothesis that taxa C, D and E have a ified from those of Greenwood et al. were published by
common most recent ancestor (i.e., a correct natural grouping); Nelson ( 1976, 1984). The order of the latter publication
paraphyletic group-an incorrect hypothesis that taxon E is largely followed in this book. The sequence of listing
does not share a most recent common ancestor with taxa C and
of groups within a higher classification of organisms,
D and thus E omitted from group; polyphyletic group-an
incorrect hypothesis that taxa A and E have a most common
once relationships are hypothesized among the groups
recent ancestor when, in fact, three most recent common to detennine proximity in the listing, is generally from
ancestors are not shared. what is deemed evolutionarily most primitive to most

56 The Fishes of Tennessee

derived (e.g., from the primitive sturgeons to the "high computer-assisted multivariate analyses such as princi
er" perciform fishes). pal components or discriminant functions, are termed
morphometries and have been of some utility in distin
Biogeography. Studies attempting to discern the distri guishing closely similar forms (see Bookstein et aI.,
butional history of lineages are termed "biogeography." 1985; Strauss and Bond, 1990). Increasingly, video
Geographical distributions of these lineages are consid camera images of fishes are fed directly into a computer
ered in light of relationships hypothesized from system and are then digitized in Cartesian coordinates to fa
atics studies to infer the distributional history of cilitate easy designation of landmarks for instantaneous
ancestral forms and possible causal factors, such as distance and angle measurements (e.g. , Ehlinger,
geologic or climatic events, for their subsequent specia 199 1). In addition, more qualitative attributes, such as
tion. For instance, closely related fishes in adjacent color patterns, fin configurations, skeletal and muscular
drainages may have had a continuously distributed com features, and distributions of breeding tubercles may be
mon ancestor occurring at a time when both drainages utilized in the analysis of differences and similarities
were interconnected; subsequent isolation of the drain among fishes under study.
ages and division of the ancestral population resulted in Skeletal (osteological) and muscular (myological)
speciation into two forms (vicariant event). Or the an studies have been extremely important over the years,
cestor may have initially occupied only one drainage particularly in the study of higher (intergeneric, inter
and subsequently entered the other through some tem familial etc.) relationships among fishes. Dried skele
porary connection where it evolved into a second spe tons were used in early studies and are still used to a
cies (dispersal event). These scenarios are often debated lesser extent for certain information, particularly from
among biogeographers and, in the cases of freshwater larger specimens. These are prepared by either boiling
fishes, the distinction between them can sometimes be and picking flesh from the skeleton, a very time
unclear. Problems with interpreting the distributional consuming technique, or by allowing bacterial decom
history of North America's freshwater fishes are made position or carrion-eating organisms to clean the bones.
abundantly clear in the several works published in Large specimens may be left out-of-doors, covered by
Hocutt and Wiley ( 1986) and those of Tennessee in par hardware cloth or similar arrangement to prevent re
ticular in Starnes and Etnier (1986). moval by larger scavengers, where insects will clean the
To learn more about biogeography in general, and the skeleton over a period of several weeks. Institutions
evolution of the many perceptions and theories associ that prepare large numbers of skeletons maintain colo
ated with this science, readers should consult the works nies of dermestid beetles whose larvae perform this
of Darlington (1957), Simpson ( 1965), Udvardy (1969), function. Also, see Mayden and Wiley ( 1984) for an
Pielou ( 1979), Nelson and Platnik ( 1981), Nelson and additional method of skeletal preparation. Nowadays,
Rosen ( 198 1), and papers found in the journals System most ichthyologists study specimens that have been
atic Zoology and Biogeography. cleared and stained and stored in glycerin rather than
dry skeletons. These specimens, which necessarily must
Methods for Fish Taxonomic Studies. Ichthyologists use be relatively small, are usually cleared with potassium
several approaches to study distinctions and relation hydroxide solution and have the flesh digested away
ships among fishes, including "traditional" morphologi with an enzyme, such as trypsin. The bones may be
cal methods, biochemical methods, chromosomal stained red with alizarin solution, and cartilage may be
studies, and, most recently, molecular approaches. In counterstained with alcian blue. Several papers are
morphological studies, meristic data (counts that relate available on clearing and staining techniques, including
to body segments such as scales and fin rays) and other Taylor ( 1969a), Dingerkus and Uhler (1977), and Taylor
counts (e.g., gill rakers, sensory pores and other fea and Van Dyke ( 1985). Radiographs (X rays) are also
tures) and morphometric (body measurements) data are used in osteological studies; these do not usually have
extremely important (see Counts and Measurements sufficient resolution for detailed studies but can provide
section). The ichthyologist may gather such data on information on presence or absence of structures and fa
large comparative series of specimens and subject it to a cilitate vertebral counts on series of specimens without
variety of techniques, including univariate and multi damaging them. In addition to figures herein (13,17)
variate statistics, in order to understand the variation Lagler et al. ( 1977), Bond (1979), and Caillet et al.
within and among populations. Detailed analyses of ( 1986) provided reasonably good illustrations of fish
shapes in fishes using body measurements, often using skeletons. Other guides are Gregory (1933) and Weitz-

ICHTHYOLOGY: THE SCIENCE 57

man (1962), though it should be noted that the former aI. , 1987; Hallerman and Beckman, 1988; Hillis and
has several errors. Moritz, 1990). Extremely refined data results from such
Myological studies require careful dissection to ascer studies, to the point that, based on mitochondrial DNA,
tain and compare the origins and insertions of various maternal "pedigrees" can be reconstructed for animals
muscles of fishes being studied. Winterbottom (1974) studied. Such studies will offer a tremendous new
provided a valuable guide to muscle nomenclature source of data for studies of relationships among fishes
among various groups of fishes. as well as addressing questions pertaining to hybridiza
The advent of powerful microscopes that can magni tion phenomena and population biology.
fy thousands of times, such as light transmission and Finally, some ichthyologists have conducted system
scanning electron scopes, has opened up a new area of atics studies of fishes by comparing chromosome num
investigation for fish systematists. Knowledge of the ber and morphology. In a process called "karyotyping"
microstructure of features such as scales, teeth, gill (see Denton, 1973; Blaxhall, 1975; Kligerman and
rakers, and many others, promises to add a great deal of Bloom, 1977; and especially Thorgaard and Disney,
new insight to studies of fishes. 1990), spreads of metaphase (undergoing division)
In the past two decades, protein data has become chromosomes from individual cells are prepared so that
very important to fish systematics. Analyses of varia they may be counted and compared with respect to con
tion in proteins within and among fish groups by elec figuration. Some surprising information with regard to
trophoretic techniques has become commonplace. chromosome numbers among closely related fishes has
Variation can be compared visually by the banding pat emerged from these studies. For example, some may
terns that result when extracts from fish muscle, liver, have evolved twice as many chromosomes as others
eyes, or other tissues are subjected to electric current through a process called tetraploidy. Further, unisexual
passed through a bed of starch gel or other medium species, such as those discussed herein under
treated with buffers and stained for specific proteins. Poeciliidae, may be triploids. Karyotypy, which exam
Differences in banding patterns result from differences ines only the gross morphology of chromosomes, may
in electric charges associated with different protein mol have somewhat limited utility in elucidating relation
ecules which determine mobility. Data from elec ships of fishes, but recently developed techniques to re
trophoretic studies, aside from its utility in veal banding patterns on chromosomes have some
discrimination between different "electromorphs," and promise of revealing additional variation for analyses.
thus species-level taxonomy, can be, as with mor It should be pointed out that, no matter what the
phological data, subjected to both phenetic and cladistic methods and data source utilized in systematics studies
methods to assess relationships among organisms. of fishes or other organisms, the interpretation of that
Much more information on electrophoretic techniques is data among different workers is always subject to con
available in Brewer ( 1970), Buth ( 1984), Richardson et troversy. The concept of what constitutes sufficient vari
al. ( 1986), Aebersold et al. ( 1987), Hillis and Moritz ation among populations for formal recognition as a
( 1990), and Leary and Booke (1990). species or other taxon varies among taxonomists. What
More recently, DNA techniques have entered the constitute "significant characters" on which to base
realm of systematic studies. Analysis of maternally in ideas about relationships are always in dispute. More
herited genetic material (mitochondrial DNA) or nu over, relationships hypothesized on the same body of
clear DNA from cells of fishes or other animals is now data treated by the different approaches to classification
possible through the use of restriction enzymes which (phylogenetic or other) may result in very different out
recognize specific codon sequences and, with repeatable comes. These same basic problems have arisen suc
consistency, cut DNA into fragments at these sites. cessively in morphological, chromosomal, and protein
Fragment sizes may be compared through the use of studies and are emerging in DNA studies as well.
electrophoresis to detect differences in sequences be
tween organisms. It is also possible, by recombinant
methods too detailed to discuss here, to clone and se Biological Studies of Fishes
quence the DNA fragments for further comparative
studies. DNA "probes," prepared from bacteriophages Ichthyologists who study the biology of fishes may be
or in a mechanical synthesizer, can be "hybridized" to concerned with any of several aspects, including repro
fish DNA to allow detection of variation down to the duction, early life history, age and growth characteris
level of individuals (see Wilson et aI. , 1985; Ferris and tics, population dynamics, food and feeding habits,
Berg, 1987; Gyllensten and Wilson, 1987; Moritz et predation and parasitism, habitat parameters throughout

58 The Fishes of Tennessee

the year, and other ways in which fish interact with amount of time and luck to encounter sufficient be
their environment and the community of organisms they havior for detailed observations, and many fishes spawn
live with (ecology). The results of such studies have in habitats that are of limited accessibility or visibility.
provided much valuable information cited throughout Consequently, this information is nonexistent or frag
this book's species accounts. mentary for many fishes.
Studies of reproduction include determination of Early life history studies of fishes entail a great deal
spawning season and habitat, fecundity (number of ova of tedious work with fragile organisms. With some spe
spawned by a female), courtship and actual spawning cies it is possible to gather eggs from the natural
behavior, and post-spawning behavior, such as nest spawns of known parents or to "strip" (extrude eggs by
guarding by males. Actual "reproductive success" is a massaging the abdomen) from a ripe female and ar
function of how many progeny survive to an age to con tificially fertilize them with milt extruded from a male;
tribute to production of subsequent generations and is this method is also very often used for artificial propa
thus related to "recruitment" (see discussion of popula gation of fishes in hatcheries. Often, spawning is in
tion characteristics, below). Much of what is known duced by administering gonadtropic hormones (e. g. ,
about reproduction in fishes was summarized in system pituitary extracts) to fishes nearing spawning condition.
atic fashion by Breder and Rosen (1966). Eggs are then hatched in captivity and periodically ob
Spawning season can generally be determined by col served microscopically to gather data on development
lecting individuals as the year progresses and observing times and the progressive morphology of the developing
the increasing fullness of the females as the ovaries be larvae. Rearing of larvae has proved very difficult in
come expanded with maturing eggs, a condition known some groups of fishes, limiting collection of such data.
as being gravid. Males of many species have secondary In other groups it has been necessary to collect larvae
sexual characteristics, such as breeding colors or tuber periodically over a season with fine mesh plankton nets
cles, which heighten in development. More precise esti (either towed or situated in current) and then attempt to
mates of spawning season can be obtained when determine development from a preserved series by
females are so "ripe" that eggs flow freely from the vent "backtracking," beginning with specimens large enough
with the application of a little pressure to the abdomen. to definitely identify to species and associating them
Males, too, may flow milt at this time, but sometimes with progressively younger stages. Such associations
some flow can begin well before the females reach actu facilitate larval taxomony, and together with collection
al spawning condition. For fecundity estimates, it is data, provide valuable information on habitats essential
generally necessary to dissect the ovaries of a series of to the fishes' early life history. While there has been tre
females at the beginning of the season in order to obtain mendous progress with such studies in the last two dec
an average of mature ova; fecundity generally increases ades, detailed data are still only available for a small
with the size and age of the female. These estimates are percentage of all fishes. For further information on
probably often confounded by the fact that, in species methods see Snyder (1983) and several papers pre
with extended spawning seasons, more ova may have sented in Moser et al. (1984).
matured in a given female as the season progressed. Biologists studying the age and growth characteristics
The general spawning habitat of fishes is relatively easy of fishes use a variety of techniques ranging from sim
to ascertain as they will congregate in that area, in some ple to rather sophisticated. One of the simplest methods
cases after a long migration, during the height of the is to collect large series of individuals from a popula
season. On the other hand, spawning microhabitats for tion over a short period of time and try to infer age
more specialized spawners can be much more difficult classes from histograms of length data (length frequen
to ascertain and may require lengthly observations and cy); for example, a trimodal histogram would indicate
searches for egg deposition sites. These are still not three year classes. Because populations are usually
known for many fishes that spawn in habitats with poor heavily skewed towards younger individuals, and be
accessibility. cause growth increments are often small in older fishes,
Observations of pre-spawning behavior, such as nest discerning older length (age) classes can be difficult
construction, courtship and spawning behavior, and with this method, especially in longer-lived species.
post-spawning can be relatively easily obtained for a Length-weight data is also sometimes utilized as an es
few species which are tolerant of life in aquaria, but timator of fish growth.
this information must be gotten in the natural habitat of Other methods of determining growth in fishes usu
the majority of fishes, either by stream-side observa ally involve relating growth rings in hard parts to fish
tion, snorkeling, or diving. It requires a tremendous lengths. Scales have been most commonly used for this

ICHTHYOLOGY: THE SCIENCE 59

form of analysis with measurements made by placing species not in the mainstream of economic importance.
slide-mounted scales in some type of projector and Further information on fish population studies appears
measuring the image from the focus of the scale to vari in Weatherly (1972), Weatherly and Gill (1987), and
ous annuli (see Anatomy and Function section, above). Cushing (1983); instructions on related techniques ap
The age-length relationship, based on measurements pear in several chapters of Nielsen and Johnson (1983).
from scales removed from a large series of fishes of Studies of food and feeding may involve direct obser
known lengths, facilitates a "back-calculation" pro vation of feeding behavior and food selection as well as
cedure so that lengths at certain ages can be predicted detailed analysis of stomach contents. The differences
and fishes from that population can be aged approxi in feeding behavior and specific feeding sites among
mately by simply noting their length. However, there fishes may be subtle; to ascertain them can require
are often problems with this procedure because growth many hours of observation in the natural habitat, and of
rates may not be uniform (isometric) over time or be course, some habitats do not lend themselves to this
tween body parts (thus being allometric) and, in some kind of observation at all. Stomach contents, coupled
populations, scale annuli, for various reasons, can be with detailed knowledge of the distribution of food or
difficult to discern. Other structures that have been used ganisms, such as immature aquatic insects, in the hab
for aging studies are cross-sectioned fin spines (particu itat can offer tremendous insight, not only to food
larly of catfishes), opercular bones, vertebrae, and preferences but to specific feeding habitats. Periodic
otoliths (see Anatomy and Function, above). Otoliths, coincidental sampling of both fishes and potential food
sectioned and projected on a scanning electron micro organisms in the habitat through the seasons, and more
scope, can provide extremely refined growth informa over, periodic sampling of fishes over the course of a
tion, with even daily growth rings discernible. For more day, can provide detailed data on seasonal and
information on aging of fishes consult the works of chronological trends in consumption and diet prefer
Lagler (1956), Weatherly (1972), Bagenal (1974), ences. Fishes and samples of potential food organisms
Moyle and Cech (1982), JearJd (1983), and Summerfelt are generally preserved in the field and returned to the
and Hall (1987). lab until such time as the fishes can be dissected for
Studies of popUlation characteristics include esti stomach contents and these contents, as well as the
mates of density, total population, age structure, recruit samples from the environment, sorted and identified. Of
ment of young into the population, survival, and course, an ability to sort and identify potential and actu
predator-prey density relationships, as well as studies of al food items-such as aquatic insects, mollusks, small
distribution and movements. Such studies require con fishes, or others-is critical. Potential food organisms
siderable sampling over several seasons and the analysis should be carefully sampled both qualitatively and
of much data. Sampling methods (see fish collection quantitatively to ascertain specific sites of occurrence
methods in following section) include intensive netting (to reveal where fishes are eating) and estimates of den
of an area with unbiased gear, trawl surveys, and ich sity, which can be compared with occurrence in stom
thyocide applications to a given area, and investigation achs to give indications about food selection. Stomach
of anglers' catches. Such sampling facilitates the ex contents of predatory species also, of course, are a key
trapolation of density estimates of varying validity, and source of information on predator-prey relationships. A
age analysis of the fishes caught yields data on age good introduction to these methods is found in Bowen
structure, recruitment (and thus overall reproductive (1983) and further information is given in Ivlev (1961).
success), and survival. Fishery biologists have at W hile the biology and ecology of economically im
tempted to better understand and perhaps predict trends portant species have been studied for many years, and
in fish population dynamics by modeling them with there has been some emphasis on lesser-known species
computer-assisted techniques. Tagging and marking in recent years, a tremendous amount is yet to be
(catch, tag, release, recapture) studies have been very learned. Even some of the well-studied species continue
important in gathering data on movements, as have ra to surprise us with unexplainable fluctuations in popula
dio telemetry studies in which a tiny transmitter is sur tion levels, reminding us how little we presently know
gically implanted in a fish which is then released and about their interactions with the environment; of course,
tracked. There is still a great deal to be learned about we have barely scratched the surface for many, more
the vagaries of fish populations, especially among those obscure, fishes.

60 The Fishes of Tennessee