Anguilla Rostrata) and Atlantic Anchovy (Anchoa Mitchilli)
W&M ScholarWorks
Dissertations, Theses, and Masters Projects Theses, Dissertations, & Master Projects
1958
A Comparative Study of Meristic Variation in the American Eel (Anguilla rostrata) and Atlantic Anchovy (Anchoa mitchilli)
Ernest C. Ladd College of William and Mary - Virginia Institute of Marine Science
Follow this and additional works at: https://scholarworks.wm.edu/etd
Part of the Fresh Water Studies Commons, Morphology Commons, Oceanography Commons, and the Zoology Commons
Recommended Citation Ladd, Ernest C., "A Comparative Study of Meristic Variation in the American Eel (Anguilla rostrata) and Atlantic Anchovy (Anchoa mitchilli)" (1958). Dissertations, Theses, and Masters Projects. Paper 1539617372. https://dx.doi.org/doi:10.25773/v5-dd73-y349
This Thesis is brought to you for free and open access by the Theses, Dissertations, & Master Projects at W&M ScholarWorks. It has been accepted for inclusion in Dissertations, Theses, and Masters Projects by an authorized administrator of W&M ScholarWorks. For more information, please contact [email protected]. A COMPARATIVE STUDY OF MERISTIC
VARIATION IN THE AMERICAN EEL (ANGUILLA ROSTRATA)
AND ATLANTIC ANCHOVY (ANCHOA MITCHILLI)
by
Ernest Cooiidge Ladd
A THESIS SUBMITTED IN PARTIAL FULFILLMENT
OF REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS
JJ KUM. TtiiS
COLLEGE OF WILLIAM AND MARY
1958 £>
TABLE OF CONTENTS
* C « Rays. * . 9 - * « Sublect Page ** ^ *. i - > a *• * * * « . * « « . *. / « ' * 9 ♦ -5 INTRODUCTION...... 1 *. / *- u ; u ' W r - ''■•:• ■'*£ Economic Importance of Eel and Anchovy . . 1
~t>i t - History and Purpose of Her Is tic Studies. . 2
Purpose of Present Investigation ..... 4
Heristic Variation in the American Eel . . • • • • • © • ‘5
Meristic Variation in the Atlantic Anchovy •■ a-, a • • • • 6
MATERIALS AND METHODS ...... 7
&: STATISTICAL ANALYSES. .' . vjf. .rv .if.?id. . y- 1% I • * • • • » 8 . ., -. tr vi ■, * *. ■ *. & « * i * <*? *. ■** * w > * - s * 7 THE AMERICAN EEL...... 1 • t # 10 r. •' - « • * < - - * » -V ■ v » v « * - * ■* . * to **- r .{ i £ Numbers of Vertebrae ...... • • * • • * • 10
■ f y *‘, V' -'if“» b . ,iaD. r- • «.-»'**' * » V V »• « * * to . ■£* > « '* « a- ■ ■ .* ■ i
Variation in Vertebral Counts with Length. • • • • • • • 10
— Precaudal and Caudal Vertebrae . ♦ . . . . « « « • • * a 11
Structural Variations...... •# • • • • • ♦ 12
> Numbers of Fin Rays...... * • • • • • * 14
Relationship Between Numbers of Body Segments and Fin Rays • • . • . 15
THE ATLANTIC ANCHOVY...... 15
Numbers of Vertebrae • • • * • • • 15
Variation in Vertebral Counts with Length. ♦ • ♦ • • .' • 16
Precaudal and Caudal Vertebrae ...... • • a • • f •16
Structural Variations...... • • 17
Numbers of Fin Rays...... 18
Dorsal Fin Rays . , ...... 18
mrr Z5£e
Anal Fin Rays...... 18
Pectoral Fin Rays...... 19
Relationship Between Numbers of Body Segments * ' vm* r,." > 4 and Fin Rays ...... „ .... 19
COMPARISON OF MERISTIC VARIATION IN EEL AND ANCHOVY...... 20 K-; .} i „ w ■: » •'< * «* » Numbeis of Vertebrae ...... 20 1 *’ * ... *J - * ‘■J *. f . t‘, . I ,. '■- *-e ' \ ? y Variation in Vertebral Counts with Length...... ••• 21 H ■’ ,. ■' , ... * n - *••; ii;i v, Ci^/i Structural Variations...... 21 v ■ , v..- - • ...* ..*.■?•■ x , i* Numbers of Fin Rays. . • • ...... 22
Relationship Between Numbers of Body Segments and Fin Rays ...... 23
SUMMARY ...... 24
LITERATURE CITED. . * « . . » «. . . > « . . « . . . • . -■» . . * 26
* *' ift .»;■> t « t * ■:-> C <
.■ -s1 \ ? A l t.
. •. . "5. *• J 4 > -■* . ‘i’v-'Uj . \'.y -'-.h
‘i. "* ft-1*; r ?*.■'' * L . > i-... U *" .• t ?; •.< • ’ :
k’jf. I, : • .. *ii*. "...
■, ;..T. i v v i ’ \ v • . i i 'it. * . i«v
^ V. ^ i i ■- i. - ^
•*. . *. r - ■ - ■: ?%.. t <• I: * -a . • • i . i *■«. . iv
v - L . - ■* ; c v .• t,} •. i-yiA ‘ . tjf <*>t •* * > - ACKNOWLEDGEMENfs ...... j.1 ; „ it in
i The writer wishes to express his appreciation to the many people who assisted in this study, :r 7 7 . I : l . -.1 -.--v . 'y
The sample of elvers from Nova Scotia was provided by Dr. A. W.
H. Needier and the New Hampshire sample by Professor C. Floyd Jackson,
Clarence Richards and Maurice Pare. Elvers from Virginia and all anchovies were collected with Mr. William H. Massmann of the Virginia Fisheries
Laboratory.'. 7 ... i. ■. * .5 . ■ . I'.-j rivs if.*''.a ,'’°^rCK,v
r Special thanks are due to Messrs. H. E. Stevens, B. B. Lineberry and W. W . Cloe of the Statistical Section and M. L. Winkler, Staff Photo grapher, of the American Viscose Corporation, Fredericksburg, Virginia.
The .typing was done by Mrs. Lois Grigsby, Miss Miriam Richards and Mrs.
Edith Briscoe,, and Miss Particia Conner. - ; sry d ‘s. vur- ■- „ r- 1
The writer is grateful to Dr. J. L. McHugh, Director of the
Virginia Fisheries Laboratory, for his technical advice and guidance throughout the study. The persistent confidence and help given by Dr.
McHugh and the author1s wife, Alice Ladd, are largely responsible for completion of the study. Special thanks are also due Dr. G. W. Low, Jr.,
Plant Manager, and Dr. C. W. Tasker, Technical Superintendent, of the
American Viscose Corporation, Fredericksburg, Virginia, and to the entire staff of the Virginia Fisheries Laboratory for their assistance, under standing and encouragement. a . M>i : - '-r i, >.£*..) u-’ -» . * INTRODUCTION :
Economic Importance of Eel and Anchovy v \ i r; r „■.. ■. v \ ;
? The eel has long been a food fish of secondary importance in
Chesapeake Bay and its tributaries, Hildebrand and Schroeder (1927) re ported that in the commercial catch of 1920 it ranked eleventh in qu^ptity and tenth in value* 318*000 pounds worth $33*704. Quantity taken and value have increased since 1920* but relative importance has dropped.
Anderson and Power (1955) reported that the 1955 catch in Chesapeake Bay was 873*000 pounds worth $94*000* ranked sixteenth and thirteenth* respec tively* among migratory food fishes. In 1955 in the United States 1*386*000 pounds of eels* worth $186*000* were landed. Chesapeake Bay.landings were
63 per cent by weight, and about 50 per cent by value* of this total. If eels were in greater demand in the United States* landings undoubtedly could be increased considerably, for the resource now is underexploited.
rv. The eel is considered by some to be very destructive of other fish. This belief was so strong in the latter part of the nineteenth century that a bounty was paid by the State of Maryland for their destruction.
Eels probably are no more voracious than other predatory fishes* such as striped bass* but their superficial resemblance to the parasitic lamprey may have been responsible for this strong public feeling. However* they * » i • - £ ; J # 4 O * * V J ' - do cause trouble by entering crab shedding floats and by damaging shad and other fishes in gill nets.
The importance of anchovies as food for other fishes is greater '' i • ft • t. ^ : . ; ; . . 'u than their direct importance to man. Hildebrand and Schroeder (1927) • • t ' * ■ t - • v* -• .VO ■ • ' ‘ stated that the abundance of A. mitchilli in Chesapeake Bay and its - 2 -
tributaries was second only to the silverside, Menidia roenidia. Results
of trawling surveys by the Virginia Fisheries Laboratory suggest that they
are one of the roost important forage fishes, rivalling even menhaden in
abundance, Hildebrand (1943) suggested that the anchovy was the most
important single item of food for large predatory fishes in the middle
and south Atlantic and Central America regions, especially for sea trouts,
bluefish, Spanish mackerel, bonito and dolphin,
t^e Old World, members of the family Engraulidae are preserved
,by salting, smoking, as anchovy paste, and in other ways. The smallness
of. most native anchovies prohibited their early exploitation in North
America but the quantity marketed in the past twenty years has increased
rapidly. Production by the United States and Alaska in 1955 (Anderson
\and Power, 1955) was 44,714,000 pounds worth approximately $597,000. This
catch consisted almost entirely of Engraulis mordax, the large Pacific
northern anchovy, that ranges from British Columbia to Lower California.
Anchovies also are used as bait in tuna fisheries along the west coast as
far. south as Panama Bay and in South America, where some species reach 10
to 12 inches or more (Hildebrand, 1943). ; u .f -i n; , -
* - ■ -f. ?■■*. j v'ef’chcft t * s i x History and Purpose of Meristic Studies t:- c c ■ ■ > . * '• r’ - ‘ ’■"» : / ‘ Geographical variation in fishes has interested biologists for ? ■* • op .* a ■ .. * „ . • r « i • * *1" ■' *- ' “ *• the past 60 years. David Starr Jordan, the famous American ichthyologist,
in 1892 was the first, to recognize that numbers of vertebrae in fishes * r " . r * • t '***-£ iC 1 - ' r- '»Z vary from north to south, and he related this variation to the effects
of temperature, proposing the general law that species from northern
wafers tended to have larger numbers of vertebrae than those from the - 3 - tropics. Friedrich Heincke (1898) first demonstrated clearly that this principle could be applied to local groups within a single species by showing that mean number of vertebrae in Clupea harengus, the Atlantic herring, decreased from north to south. Johannes Schmidt (1920) showed that mean numbers of vertebrae were inherited, but that they could be altered by varying temperatures during embryonic development. .r x
* , ' 1 i, J . ... V ' ' * ' t 1 < disappear. This would have a bearing upon management of the fishery, for if it were evident that the resource was being overexploited in a certain region, the management program necessary to correct the situation, if the species migrated freely throughout its entire geographic range, would be quite different from the program required to prevent over-fishing of a local population that did not move far from its place of birth.
In North America these methods reached their widest application in the work of Rounsefell (1930) and Tester (1937) on Pacific herring ■-.? . , r- .... • ~ ^ ^ (Clupea pallasii) and Hart (1933) and Clark (1936) on Pacific sardine .• ...... - .. « . ... - .** • -*•••' or pilchard (Sardinops caerulea). Both species inhabit a wide latitudinal range along the Pacific coast of North America, and both exhibit considerable 2 variation in mean numbers of vertebrae, decreasing from north to south, v. s. . - v . 1.^‘fS % v: V o*. j * ' '* > ■ i t t * • * >' i , *„l. - i n vt :i^iiiliv% Z % o x vor fccbz-it* n .v *
XX" ;.i - ;• •v'-s*- two y::'M . The Pacific northern anchovy (Engraulis mordax) occupies much
the same geographic range as the herring and sardine. Mean numbers of
vertebrae in this species also decrease from north to south, but variations
in mean numbers of dorsal, anal, and pectoral fin rays are even more
striking (McHugh, 1951). The common anchovy of the Atlantic coast
(Anchoa mitchilli) exhibits similar variation in mean vertebral number
(Hubbs and Perlmutter, 1942). - c' ^’r v '\‘A *■ *' c-
cdS.trVsts zo th e .Euro;. ta n r t ic coa-tr, l'c ■. :. -.v et Purpose of Present Investigation £.4,1,-'.,~ • V r H l.U f .* * I £4 £ % • ' Individual variation in organisms of the same species, as Darwin :;;uv ir * 4.. . v y '• * ■**; ^ and others have recognized, is the basis of evolutionary change. For this pr,( ■-xt■ ■ -sv,. -v ' : • /V - - ■. --c -- , reason, variation is of fundamental interest to biologists. Fishes are ar-'i -4 4.,. .#c- . 4 - * * -i-o x particularly useful in the study of variation, for they have evolved into ■. r. •. Xv , ‘‘S' ... - v . '{'2. • » ^ '» a large number of dissimilar groups, possess many well-defined morphological
structures that can be expressed numerically, and can be obtained in large C'ifi ;.&15 . . '4 4- . tkU- •»»»> »>• »< r t, numbers with little effort. esriiact t**. •. ->* . jy c. .. dv>.V.r..?»: *. ■ v •?•*•’» of E>v , x* o*. For a comparison of individual variation in two dissimilar fish 1 0 £ in t .no • v; tj. m*v. - ••• v- • 4 vi... •: species, the American eel (Anguilla rostrata) and the common Atlantic varxtty. anchovy (Anchoa mitchilli) were chosen. The eel is long-bodied, with ...4 ■ ... j, i. v<4 ‘ • - • 4-1 »•. »■' i < i* • *• ' about 107 body segments, and a long and continuous median fin, beginning
a% the nape, extending posteriorly to join with the caudal fin, occupying - -1'h - f ; .. .r.- the mid-ventral line of the body as far forward as the anus, and containing to dirxf ' -vs.'■•vra .1 * :•... •. nc* .. ♦ » '• «. about 440 supporting rays. The anchovy, on the other hand, has only about • £ t£ «? •: . ■ - ' % •’ - • - >. *• ‘ •• 42 body segments, and relatively short and distinct dorsal and anal fins, cun*.‘ e v :.".a c - ..A - ■ •. - • - ‘ > ■ ' - ' - with about 15 and 28 rays respectively. It was decided to investigate
similarities and differences in variability of numbers of vertebrae and
fin rays between these two species. Meristic Variation in the American Eel
The fascinating story of spawning and migrations of the European -i' and American eels, discovered by Johannes Schmidt (1914, 1916) is well known. By systematic collection and interpretation of biological and s. physical oceanographic observations he found that both species spawn in - :7 *•’3 { ; the open Atlantic, in the vicinity of the Sargasso Sea, and that their
larvae, hatching out at different depths, are carried by different sets of ■tf! '• ■>'/.: j i - ■ currents to the European and Atlantic coasts. He found also that the ; ^ i - ' ' period of larval development is nicely timed with the drift of these 5\ i currents, so that the European eel does not metamorphose from the trans- i r . parent larval stage to the pigmented elver until three years have passed, ' * ' *■" and the larvae have' arrived off the coast of Europe. The American eel, -£-v/ v ... ■' faced with a much shorter journey, undergoes metamorphosis at the end of ■ \ . ■* •- ‘ r;‘ r * ~ \ i a year, in time to swim up the coastal streams of the United States and
Canada. Schmidt was able to distinguish the two species, even at the '■"*t x 4 .ij-. «•*., • earliest stages, by the sharp difference in number of body segments, about
107 in the American species, and 114 in the more slowly-developing European .V- -hi'f.ri /• • ' S' variety. - 5 r , . . "ill. :* \ ! .. l:sl, ■ ■■ . * Although he recognized a difference in mean numbers of vertebrae >; . .■ i . - - » - ' v -r . * . ■ *" between European and American eels, Schmidt concluded that there was no regular difference with latitude in groups of American eels that migrated r t - »■> to different rivers along the Atlantic Coast (Table 1).( Applying modern f. I- statistical methods to the analysis, however, there is only about one ; ■ z . * * , chance in 1 0 0 that his five samples could have been drawn from the same > j ; ~ population (F ~ 3.24, Fq ^q I = 3.34). We know from other investigations O CM rH CM CO rH •• ^ * • • rH pH pH rH rH
CO G O • o vO CM < t O i rH - “ U ; . ’ X •M vO 53 00 OO' 00 O G* pH m rH O' 111. A„ . < * V (H ' c >, a-' v &. sj * ■~xk.£‘ i ( d i O Jo uh CO CM r ^ r^. 00 00 c ■ pH , < t - 00 • 00 t v ■ CO o CO .. O ' CM ' ' CM • I - -r.bCY oa a> « • ' • • • u S vO uh - * O o O o O . i ,j t** A fte r *h.i * w -■ rH rH • rH rH rH G5 ** 4-i G Q . . . i ♦ ' - <3 M • CO • G> pH rH sssmi;*?;®,-. «S . ■v ■ * \ * &■ ; v i- ?■ Pjvi,- eSilUjslnev i. t X ■?,.- ' ' ! * V •l § «j h* O rH O ' rH r>. co 4J rH • «-* *'■
Schmidt * s data, except to suggest that mean number of vertebrae may vary between yearbroods or that eels become segregated on the basis of some characteristic that varies in regular fashion with vertebral number after they reach coastal waters. Errors in counting also might be expected but this should increase the variances. Variances of Schmidt's samples are much more uniform than for samples examined in the present investigation, but when his samples are combined and compared with f combined samples examined recently, variances are .identical. ^ ■ ' ■■■"': *■;-- - ei -• t\- Meristic Variation in the Atlantic Anchovy |
Bubbs and Perlmutter (1942)have shown that mean numbers of vertebrae in Anchoa mitchilli vary from approximately 42.0 in Mass achusetts and New York to 39.5 in Florida, Texas and Yucatan. Hildebrand
(1943) stated that two subspecies exist, A. mitchilli mitchilli from * - Massachusetts to North Carolina, and A. mitchilli diaphana from South
Carolina to Yucatan. Data from Hildebrand's table 18 for the northern subspecies (Table 2) treated by analysis of variance, gave odds of much less than one in 1 0 0 that the nine samples could have been drawn from a homogeneous population (F ~ 5.18, Fq#q^ = 2.54). This appears to demon strate that the Atlantic anchovy, like many other clupeoid fishes, is segregated into local populations along the Atlantic coast, J f * * ^ j* • ^ TABLE 2 ^
Frequency Distributions of Total Numbers of Vertebrae in Anchoa Mitchilli mitchilli, Means, Sums of Squares rd 4W tJ X> a *H 5 u Q> £ U 4J a K to
U •u O C tn 3 m c
»X vh O to st X t CM CM vO 00 00 st r^.O m c M M H H N H CM CM t f t t ^ St n s o os \o os cn m n i rH CM CM XT •d -d r—4 w m c • to (Q s o rs oo o os vO rH CM \D OS 00 o s f s n i n i H - o «-l rH h t s f . o o co co to n. O H •P 4J • s 4> P 0> & n i rH vO so r^. 00 CM m rH « « § *H o CD 00 rH vO r>« in CO o cn so O CO lT> CO Sf Sf t < s * o r*^ os m » H r •d S U. >S h > - 7 - MATERIALS AND METHODS Samples of unpigmented elvers were collected in minnow seines or drift nets in the spring of 1951 from Nova Scotia, New Hampshire and several locations within Chesapeake Bay and its tributaries. Anchovies were collected in a surface trawl (Massmann, Ladd and McCutcheon, 1952) from the James River, York Spit and the Rappahannock River in Virginia; and at Elk Point, Maryland, near the head of Chesapeake Bay. | Samples were preserved in 10 per cent formalin. For accurate i.. _ .__ _ enumeration of bony body parts, specimens were processed by a modification of the staining method described by Taning (1944). After rinsing in Tap water, they were placed in 4 per cent potassium hydroxide until cleared. This required about 18 to 20 hours for both species. Next, sufficient acidic alizarin was added to color the potassium hydroxide a deep red. After~ 8 to 12 hours in staining solution, the following transfers were made; (1) 70 per cent KOH, 30 per cent glycerin, (2) 40 per cent KOH, 60 per cent glycerin, (3) 10 per cent KOH, 90 per cent glycerin, and (4 ) 1 0 0 per cent glycerin to which a crystal of thymol was added as a mold inhibitor. Schmidt (1914) used an alizarin method developed by A. Strubberg. All meristic counts were made under a dissecting microscope at appropriate magnifications. An eyepiece was fitted with an insect pin to serve as a pointer. Figure 1 shows the two species as prepared for 'V ^ * meristic counts. Eel vertebrae were counted starting with the atlas and ending with the last hour-glass-shaped centrum, thus including one less vertebral *.x ■. 1': • I a i >. i ATLANTIC ANCHOVY AMERICAN EEL v"% *> fy t> ^ / S / f / ■' /■'/■■/ ' ' . V •' r / >‘/ / / J ' y . v- * v -'-t j ^/L, <*■**-'"■■’■ . ■• v v > , - . V . \ \ \ A v^ N V'x V v ' A 'A \A \\ * 4 - V' '.v- I'. '• •. ' . .-- A ' Ai * A \ \ \ i /}\ ^ 2 !«•-«%* fa&H \ \ \ { \ i '• ; • - • • ^ § 1 L *,*’ ■. *\ --V. \' \^^-\jU=4iS5mgamimSB^ **$ «?**??£*? ** *&£#'$&?•’r^* ' ' * / * - / ^ / j i i fs?r?|?$?* ," / «*£»*?$* / < $• ,1*^_T^3iAw«u,X• *i*.- \>%...id*** Figure 1. The anchovy and eel as prepared for making meristic counts, showing vertebral centra and fin rays. - 8 - segment than was counted by Schmidt (1914)* A count of total vertebrae was made on 693 specimens; 298 from Nova Scotia, 296 from New Hampshire and 99 from Virginia. Counts of precaudal vertebrae, caudal vertebrae, neural and haemal spines, and dorsal, anal, caudal and total fin rays were made on 100 specimens from each location* Standard length, the distance from tip of snout to end of the hypural plate, was recorded for all individuals* J'ur- nc.srl? i.li b'loXc«£t,*ni c ,ck. d faarrL'^ft The following characters were counted or measured on 704 anchovies from five Chesapeake Bay locations: , precaudal, caudal and total vertebrae; neural and haemal spines; dorsal, anal, and pectoral fin rays; standard length. Vertebral counts of anchovies were made by a method similar to ’u* that described for eels, starting with the atlas and ending with the centrum just anterior to the urostyle. O.. : — J_' Sketches were made of deviations from normal structureof vertebral centra or their processes.! If there was indecision regarding the proper count when fused centra were encountered, the specimen was not used. ./rs*#" to. the chexectern voder iucdr- I - rioce heterOgvTieitv * is "b-' h» ••d F or * Va &.: ur*; tv. : Avi:.o* * ; STATISTICAL ANALYSES coast, : . i - - •< j-.'?-'" it * >•••■* r *' In analysing observational data such as this, two quite different t ' .1 A.. • .'"-t . ■ t -■ •- I-.. ••• ■ methods have been employed by biologists. The first is to arrange the information in some systematic fashion, so that characteristics can be •Oi “• . -- ‘ .. • - - -■ -■-*. *• v • ■ - • • * compared sample by sample, and a subjective interpretation may be made of any difference that appears. If the information can be expressed in numerical terms it is often compared graphically. This method, still commonly used in biology, especially by taxonomists, permits no estimate of the probability that the difference is real. The second method begins with an hypothesis, the truth of which can be tested by examining samples and treating the data by statistical methods. If conducted properly, this method removes the possibility of personal bias. It is not intended to infer that the first method always gives incorrect results, for if the observed differences are sufficiently clear-cut, and especially if there is no overlap, statistical treatment is not necessary. But since nearly all biological observations are based on samples rather than on the entire population, and since estimates based on Samples always are subject to error, it usually is advantageous to be able to state the odds that the decision is correct, ^hen differences between samples are relatively small, as they proved to be in the study reported here, statistical analysis is necessary, for there is no other way to test the hypothesis that the differences are not real. - , : rep »' Although the primary purpose of this study was to compare individual variation in the two species, it was also of interest to test the hypothesis that individual samples of each species were drawn from populations homogeneous with respect to the characters under investigation, since heterogeneity has been demonstrated for the anchovy along the Atlantic coast, and since there is some heterogeneity also in Schmidt*s data for the eel. In the comparative study of variation, however, it is immaterial whether the individual species were drawn from homogeneous populations or TO-'; J i ■ , V*# y- , , * . t \ 't 1 not. Therefore, details of statistical analysis have been eliminated in "*v * i , f*. ■ i* U . " iv. ,"*iy V •.'*'? V ' ”* • * * * f - * •' discussing heterogeneity within species, to avoid excessive use of V&\ • /.** . -v-. - I *- j -' in " *~.z. *'*&.. -v*» - . ’ statistical terminology. .c l z tecfietie • otvdi-.- ch . *fcV h-i r i • - 10 - Absolute variability is greatest when the characteristic being measured is large. Thus, the variance of body parts in two species as dissimilar as the eel and anchovy cannot be compared directly. To permit comparison, a statistic called the coefficient of variation (C) is introduced, which expresses the standard deviation as a percentage of the mean: C = 1 0 0 s. M ; Jl The American Eel (Anguilla gostrata) - ^ ’ i Numbers of Vertebrae ^ -t-? •- ’ - ^ - Total numbers of vertebral centra were counted for 693 elvers (Table 3). Analysis of variance gave no indication that- the samples could have been drawn from different populations. As mentioned already, these means compare closely, in magnitude and variance, with those reported by Schmidt (Table 1). The range in total counts of individual fish was considerable, however, for the highest counts were about eight per cent greater than the lowest, but the coefficient of variation was consistent with that calculated from Schmidt's data, only a little more than 1 per cent of the mean. Variation in Vertebral Counts With Length It has been shown for Pacific anchovy (McHugh, 1951), Atlantic herring (Ford, 1933), and various other species that numbers of vertebrae within a fish sample may vary in regular fashion with body length. Various reasons have been advanced to explain the phenomenon, but the knowledge is important in meristic studies chiefly because fish often are segregated in estuarine waters according to size, hence variations in »-: ' v« / •- *. j V.4* Cv.r*. t i; < ; .CJ. XL , * 1'. tT i 5v>. .*■ ..‘ij. '«cr ■ ■* * •o- , t J ■ l?* : u: ■Ji”'-!;. 1 real Che : he »•;.: lo c r’l ' fci ‘ t tv , 5 • « v -» n, ? ?•<£, jo ‘r rSiS, t t d Vi: r n i & * I - i t - z: s.:-C€ a*'. t« U .4 . no i. * *.ctS; 4 r ; ';■. Vv- .: «uf hr.t ^Itc- v^vr>. ; &z.i .r: ; * •.. ~<• .1 * 1ft '-tt ?arc ; *r ■• . i.'!-",y.t. ii h ‘> • \> tv t : v<* - ,i~ TABLE 3 •C *» • ■* , ' *. $ *. I <..*• i'li'Ktr-P .■■ . a '*:€* : j.tn .;i7.' t* r.-'/vj -j*.c it *• Frequency Distributions of Total Numbers of Vertebral Centra in Anguilla rostrata, Means, Suras of Squares of Deviations from the Mean, Numbers of Specimens and Coefficients of Variation from Three Localities, •*•.*. • *. fc i’V .*1,4 3L*i tt;.i VT ' nu«- f-a.■■■;■. >t.- i- - : Numbers of Fish Having Total Vertebral Counts of: Locality 103 104 105 106 107 108 109 Mean S (x-x) ^ No. C Nova Scotia . 9 > 2 2 55 96 72 .35 9 106.14 510.79 298 1 . 2 New Hampshire 6 15 64 104 71 24 12 106.15 446.75 296 1 . 2 ‘" ' * L" 3 Virginia 2 19 42 22 11 106.13 113.29 99 1 . 0 c tc.'sr a vten ot wrrtcbra 'v..th -s. in jj>; j ;#‘t ». .ii. at b I ■& * 0 L,n £& CimCli V-iT.CS • <*se *■ T- 4* bter '1937^ £t-und i ru.t \ * : x,; 1 i •;■ i- *: Jff" 4 c ;-vr.: ;> ci jL" ^ *-*' i 1 / - i.-. - v- b e . ^ ee u 1 on n 1 it i *&?* ues .cor ’'it*i'-,4 ■ ^ ’ V • fc- ?' , * * *. , * *?*. c i chii -Ir aicb ne, i4c.-l>2,h -v194i.) found t;V i:,: i r * ■t: _V/i* V r J . ;£rv»‘ 1.1 c ,£ 1.K ;rin; ** 5? • f’ h* hi j o o r»: e t u. •*v ? • ■ *. * * * ' -v*. i*nd. . e s k * y cort:«2 .r, i- t ii* f I i l' ,V. h & I ; efeo,- t h a t -uo ss.g--.ifi cant d : t .V.TC !• .. t 'i - i H .. lsj .rr:i*i iv'js th.* it . ,|o; \j.*\ „** I- X „i * l is,* ^rcr*'' ti-‘ :tir ^ XL tc uTttT i-; t i * j.ia? - IX - mean numbers of vertebrae, and perhaps other characters, may be merely an expression of this differential schooling. Average lengths of elvers from the three localities were: Nova Scotia, 57.5 mm; New Hampshire, 56.4 mm; and Virginia, 54.5 ram. Differences between localities in total vertebral numbers also were insignificant. The total range in standard length of individual fish was 48 to 6 6 milli meters. Treating the three samples as one, mean numbers of vertebrae and sums of squares of deviations from the means were calculated for fish 48 to 57 mm. and 58 to 6 6 mm. in standard length (Table 4). The difference of 0.39 ^vertebrae could occur by chance less than once in 100 trials. Therefore, to interpret observed differences in mean vertebral numbers of elvers, it would be necessary to know something about their geographical distribution } • , ; - - < : with sise. It is possible that the differences between Schmidt*s mean vertebral counts (Table 1) were caused by such a phenomenon. -Tr fe*. ^ ‘ ♦ I-\ 4«c' * "*' ■ * T % k * » , * f . b' } I Precaudal and Caudal Vertebrae *' ^ ’ : v Anterior vertebrae with haemal arches open are known as abdominal or precaudal vertebrae, posterior vertebra with closed haemal arches are known as caudal vertebrae. Tester (1937) found that variation in vertebral counts of adult Pacific herring between localities was confined to the abdominal portion of the backbone. McHugh (1942) found that total vertebral counts in young Pacific herring were highly correlated with abdominal vertebral counts and only weakly correlated with caudal counts. It has been shown already that no significant difference exists between mean total vertebral counts in elvers from the three localities sampled, but it was of interest to determine if counts in particular t •• r-5* "i's'i iH 1 •■’bi i ••• V*-, •« ' 2. n ^ * v? r a t „ ' ;h ' -pr*-: «\.r • ;Kd loir •* It • ic •:• ♦. '^i **-; wnlfiej, i f»‘ •■ ■■■••■»v : • ; i.‘. o»; ec*.-nti» ir« fcpe t *' ri'"io: 1f h? bo,* j *. : t. , ''C..vSt.» ct ' ixi-.'t U*-: ,• c • ; i •.:; t* i t . TABLE 4 *. K Variation in Mean Numbers of Vertebrae with Length in Combined Samples • -I’C.: i; > :, , a., ‘ * ' • ‘l •>*. .-*d of Elvers, Anguilla rostrata. J 5 19" i.) . i ; Length mm. Mean S(x-x)^ No. c 48 - 57 105.98 37.70 407 0.9 58 > 66 106.37 21.23 286 0.8 * ; L u:,; v f •' ;vi •- . • . - *‘.K ' - .. - ■ n L>; •, <. 'IV; M id I t . .‘ 'Sit i 'i of $ & . - •' tS‘^ rk _.v«? f-. - e i f-lvers * %i 1 ».• i C- a * ? "i. • ^ »i - , t . r . c h^T. *'i t. i * .« 2: * 3 ai'.d sp i ceil v ar<‘-.”'ysv c‘‘■i ; v '. Ifi* * t Z* * J fr'iyu*:-.- 2 ) . :•;? rt t v j M .*’*4 • *. mf i 1 k * i 5 - ^ 4 /;*;■ t f i r;C itt 36 p-2t r.;*»£ of all <«4l .•<: **.. ■: ; A. .1 t.-c : u v t. *»''r:,; Jtotis oe : '■*? • y it* *. c.u !* rsf*.-* ’ r ' ■* ti ?■' li, fe-oc^rdiag C-o fc-q/ti f .*.v 13) a.ut .-.-cHt .;;h , i 9 ‘- , *;• Ari.*'£ f lv V *ii.* U'-oy &?<*.'itru&i l / is tVe •**-*i, nearly Vo pi?/ ~ M i a . -o ^st'vlcr vhycofcLVy. regions of the backbone were more variable geographically. Mean numbers of precaudal and caudal vertebrae in elvers from three different localities differed by slightly greater amounts than did total counts (Tables 5 and 6 ), but the differences were not highly significant statistically (P 0.05). Relative variabilities of counts in the two regions of the backbone, as expressed by the coefficients of variation, also were similar. ‘ 'A*. Structural Variations ;:*.•»& a t r f I - h : -Jy-K ^ la Considerable variation was observed in structure of vertebral A.- .» " i> c •. 1 'i: ■ > :!.i/ columns of eels. Variations in structure of centra have been described ':lv * * ‘ - i * i ‘ C ?. •* '< A•'*' t Ok, , -f by Schmidt (1921a), Ford and Bull (1926), Kandler (1932), Ford (1933), McHugh (1942), and Gabriel (1944). Their interest in such "abnormal ities" arose from difficulties in ascertaining true numbers of vertebrae in backbone in which such variations appeared.. Only seven of 700 elvers — • . < ” - 5t r' 1 had abnormalities in centrum structure that led to doubt as to the correct i . 1 ; ? count, but 397 individuals had one or more minor "abnormalities". Table 7 : . i . i gives type, frequency, and location of skeletal "abnormalities" for each of the three samples of elvers. Modifications and duplications of neural and haemal arches and spines were most frequent, and represented 95 per cent of all "abnormalities" (Figure 2). These could hardly be called abnormal since they were present in 56 per cent of all individuals examined. These types of structural variations occur frequently in connection with "normal" i vertebral centra and, according to Ford (1933) and McHugh (1942), in Atlantic and Pacific herring they are usually associated with the last two vertebrae preceding the urostyle. In the eel, nearly 76 per cent of all "abnormalities" were confined to the two posterior vertebrae. TABLE 5 “~r , i Frequency Distributions of Numbers of Precaudal Vertebral Centra in / 1 Anguilla rostrata, Means, Sums of Squares of Deviations from Means, Numbers of Specimens, and Coefficients of Variation from Three I1* i • Localities. V* ] **' **' ' C3. I* i . " Numbers of Fish Having Pre caudal Vertebral Counts of: Locality 41 42 43 44 45 46 Mean S(x-x> No. C Nova Scotia 18 96 143 37 4 .. . 42.71 195.61 298 1.9 New Hampshire 2 0 83 156 39 2 1 ... 42 • 84 151.28 286 1.7 v • [ Virginia 4 25 55 16; . -V 42.83 54.11 1 0 0 1.7 TABLE 6 ...... > - » e '» 53 V !‘ •M CM -a r—i H o «Q O 3CO *M .3 *H u t < ftt Hta Q •H 4J 4J CO cO 0O <0 u <0 U *3 ■ J % -■’ _* * ‘ •_* -■ ’ ■ J3 m c S *r4 33 •M f*4 525 (0 tO O UD <0 O 3 U > 3 B CM Si r-4 v0 T5 r-S O U > 4J 4J CO CO O 3 to • • 3 cd o M <0 U 3 CM 1* N_> /"S vD vO M5 in CO O NO 10 525 NO V© s n O00 CO r*- M CM to o M• rS3 •H ♦M - U r-4 M cd o O 691 , O o ro © ’ A z; co © © 1 5 4 co h4 u O © - d • © e o o o o o 3 ss o o o o § • .25. CO CO r-4 r- © •. n s ft. d © •o d •s © © u © t~i O © U 3 O © u X oq CO 00 © © r-4 © 4J < Cd © M B u > U © 3 © u-4 o o ts > o 25 H <1 F X \ X Z X Figure 2. Diagramatic representation of (A) "normal'’ and (B-H) various types of ’’abnormal" neural spines in the last three vertebrae preceding the urostyle• Similar varia* tions also occur in haemal spines. r - 13 - The number of "abnormalities" per skeleton ranged from one to 41 as shown by the following frequency distribution: '• “ T 1 vv■ (A) 0- 1 2 3 4 5 6 7 8 ; 9 10 11 12 13 14 15 16 17 41 (B) 296 194 69 30 17 13 11 15 14 12 6 6 6 3 1 ; 1 4 1 1 -a.vj in (A) ss Number of abnormalities per vertebral column * inn o • 15 ~ (B) sr Number of vertebral columns If.Z-lXtr , > . ; _ w, r, - n * Usually only one "abnormality" appeared in a vertebral column, but ■, . .• , - , , .. occasionally larger segments of the backbone were involved. Anomalies involving two centra were fairly common, but less frequent, and the 1 C : • i ~ - . . : * <. k *• :t . ' 5 . . v. „ ■ i „ .* -v. * frequency of "abnormalities" affecting more than one vertebra decreased steadily as the numbers of vertebrae affected increased. Distributions of this type, in which mean values are low and frequencies decrease pro - u- <• ; - • ' ; 1. L l* . • -■< .* *• ij rapidly to the right, are known as Poisson frequency distributions. They are common in many biological phenomena, such as counts of oyster strike, W i - < 'W ^ *■ t ' ' * * « r p f - * ' ' , * *" * ^ K «£ ‘ j 1 ' * * where the mean may be relatively small and large deviations to the right ^ --‘ * V .•/-,« CT-iZ^ i - v . i. i >*•-. C ,4 -> - * . ' i.:- • «, of the mean may be more probable. P k &* ""“Si? 1, ^ ^ * As demonstrated by Ford (1933) for European herring and McHugh 1.(1942) for Pacific herring, the mean for "abnormal" vertebral columns was approximately half a vertebra lower., Schmidt (1921a) and Ford and Bull (1926) concluded that the difference was caused by fusion of two adjacent vertebrae, and suggested that this reflected a tendency for individual fish to possess fractional total numbers of vertebrae. Ford and Bull (1926) Prestored" the "normal" count by adding 0 . 5 to the mean number of "abnormal" vertebrae, but it is difficult to see the merit of this adjustment. McHugh (1942) pointed out that "abnormality" was a purely subjective concept, and it is obvious that a structural modification that occurs in 56 per cent of eels in a sample cannot be considered truly abnormal. "Cl&RARV of inn ymO^lA H^HSStES i^SORATO^L - 1 4 - ■ j Table 8 gives the frequency distribution of numbers of vertebrae for ’’normal11 eel skeletons and those having abnormalities in the vertebral processes. The differences exceeded their own standard errors by factors of 3.14 in Nova Scotia (P< 0.01); 2.52 in New Hampshire (P « 0.01); and 2.03 in Virginia (P = 0.02), and therefore, were highly significant statis tically. Analysis of variance gave odds of less than one in 100 that the difference between ’’normal” and "abnormal” means could occur in sampling a population where no such difference occurred. There also was no evidence that locality means were derived from a heterogeneous population. It seems clear, therefore, that although so-called ”abnormal” vertebral columns give lower A mean vertebral counts than ’’normal” skeletons, the discrepancy should produce no bias in mean vertebral number. Ratios of ’’normal” to "abnormal” i . v t ' : backbones appear to be constant, regardless of locality, hence their in fluence upon mean counts need not be considered in measuring latitudinal variation, as long as counts derived from such fish are made according to -if Vi -M , pre-established criteria. | "-s*? Numbers of Fin Rays | * u, K. . -j * j * i ■ ■■ -.r»s r*% r- * Counts were made of dorsal, anal, caudal and total fin rays for the first 100 specimens from each of the three localities. Where injuries or other abnormalities appeared to affect accuracy, counts were not used * t in the calculations. Counts of caudal rays were made because this fin is continuous with the other unpaired fins, and it was difficult to recognize the distinction between the three fins. * * CM CM CM O • • pH pH to ft) (8 M O (8 0 3 »~) -,■•>• cr co ft) pH n% VO o vO cn ft) CM rs cn vo cn i£> cm U pH p H pH pH J*„.;ul> 2 - cr. -v* o xs H (0 1 he rr.» v'■ *5*&cue to '3 UH «\ CM n;: t ' : tlii.;- i r % » 3 /H v& o- m s t c o lO m Frequency distributions of numbers of dorsal, anal and total (dorsal + anal + caudal) rays are presented in tables 9, 10 and 11. t Analysis of variance gave no evidence that samples were drawn from a heterogeneous population with respect to these characters. Though some of the differences seem large, they are insignificant in relation to the great variability of individual counts. The coefficient of variation (C), which expresses variability in relative terms, was somewhat less for total counts and this was not unexpected in view of the difficulty of distinguishing dorsal and anal from caudal rays. i. ?. r ; ] i-* | hr/ fs. r** u% 1 .. • N- ' f Relationship Between Numbers of Body Segments and Fin Rays Casual observation suggested that variations in numbers of dorsal and anal fin rays were associated with variations in numbers of body seg ments. Correlation analysis confirmed this hypothesis (Table 12). These fins, especially the dorsal, are long and extend over most of the body length, hence occupy linear distances similar to those occupied by the . ■ * • vertebral column. There appear to be two rays in each myomere, or two v. • , . * • n rays for each vertebral centrum. Because the fins do occupy almost all available space, it is easy to understand, on physical grounds alone, how variations in numbers of rays are linked With variations in numbers of body segments. 4 . - ?v | vf •(* •" . j. ! » r»*- * i' ~ r' -X I" ! . > ’ S r ) ■* ; * : THE. ATLANTIC ANCHOVY — .. .. :s .’n' r“ 1... ■; ’• . I - i ‘ i _Numbers — of . Vertebrae. . ^T * * ^ jf Frequency distributions of total numbers of vertebrae in eight ' f * V samples of Anchoa mitchilli, 704 specimens in all, are given in table 13. Analysis of variance showed statistically significant heterogeneity at Q O r-4 CM • • - • i n m m • X) O VO o 3 25 a v o ON f e r-4 o p * 5 O M 3 CM o m o IM O /— r—4 r-» o IX • • • W CO . I CO 00 3 3 X O cO H o t-4 w O 00 o •rl o CO •s *» •V J~S 3 CO CO CO 3 V r-4 i—4 r-4 ■H P > c r 3 ■■—» Cx m O f t M 3 CM o o Pn 3 • •• *M 3 CO CM o O 2 CO CO CO • CM CM CM « m 0) 0 M •rt 3 4J P •H ,» . . -A a O' i—1 00 CM ' ’.I CO. CO 3 u*> O CM *W O O r4 CO r-> r-4 3 0 m g 0 CM P U CO xs H 00 4s- . T>». i n •>1 o - CO E CM a o • » 3 *44 0 *w O CO r-4 r-4 i—4 2 * 60 Ht i—1 r-4 r-4 3 3 3 CM •n 0 •rf 4J 3 *r4 > 3 4J •M 3 P oo Mf? O', 3 3 S3 0 co r-4 r-4 U •W - u CM •M M X* Os CO 3 3 > o > •rl 3 CO CO CM U t*< od CO r-4 CM r-4 . *4-4 1 CM 3 O CM r-4 H O 3 H r-t 3 3 00 VO O 00 •rl 4J M M CM »H. i—4 r-4 P 3 3 O CM 6( 3 tO 3) 3 •r-4 6 Frequency Distributions of Total Unpaired Fin Rays in Anguilla rostrata. Means, Sums of Squares of Deviations from the Mean, Numbers of Specimens, and Coefficients of Variation from Three Localities. Frequencies Grouped by Intervals of Fifteen Rays. — • - Numbers of Fish Having Mean Total-Ray Counts of: Locality 405 420 435 450 465 480 495 Mean No. C Nova Scotia 5 17 25 21 22 4 2 443.70 37,812.25 96 4.5 New u.-'Ver I Hampshire 8 17 21 36 1 441.25 36,468.75 98 4.4 1 1 4 Virginia 8 - 22. 28 17 .. 12 4 437.93 33,257.70 92 4.4 . 1 i o — 1' 1-* --*»* '■< ‘ t'S'— , ft, * JV-*^ .* ^ ^ . E'-'i *** ' <*** f V 4 r*. v£ r-.’ r S -if** ; TABLE 12 >s Correlation Coefficients and Levels of Statistical Significance ~ Between Dorsal and Anal •'Ray Counts and Total Vertebrae in Anguilla *. r. o # ■*** i . ^ fy \ fZ.} ^.T rostrata. • j ^ « 4 * ' r ’ o o o ^ o ^ ■4* v f •I 1 •V ' T - •- I S t i * * * * f ; Correlation Characters Coefficient Probability Dorsal and Anal Rays - 0.732 ^ < 0 . 0 1 Dorsal Rays and Total Vertebrae • 0.781 ~ : < 0 . 0 1 Anal Rays and Total Vertebrae 0.757 •- < 0 . 0 1 * o H-* Total Rays and Total Vertebrae 0.754 • o /V "> > • *■ ■% * - r. t v. * * o -•>. V » v*„ 5 ''•* ‘ ‘ . ? *-L.v ^ y *•’»* * 4' ■£ ^ 5»> f f} r4. V*v > *5 5>"t t n vr. C*v t.n C/- CV ? «“«! t-*< " ^ ■i ri s -•4 & ■ ■*.» * * iT • • <■* • « *■ 4>- tT ■* a * -*j; T .& ■ i a. a *; . -«/. ’.j [ V-% .4 esj V ^ v, -*• r/*. o E * 44 4f c n v o »r» vO in O O d r-4 at a> £ at J3 vi (0 o vO O o o CM vO o o 4t O o o f'*. 00 o i ,, .r ;■; O jd (4 H ■■..-.( >4 60 iw -r4 f c t * ^ i - - f i « . r d * o o -r4 (W tc •> :r > . tf *• < vt , ?-rf, * ■ ■:._ e - < to d / 4 VO cn ov in m 00 cn o •H o *x r-4 cn cn r*. r-. CM CM o > *r4 1 • • • • •• •• 44 .?' ■' 't *>'"} X t-4 «n r- cm vO , H f—< -v.-viu*'.; -:g at w CM r-4 CM Mt- cn cn cn •*4 05 iw M o at - , o< ; ■ i-1 a 2 > a/- C.v 1: •* i .-.g i t * 0) ■&• - -p : . .wv i Vs . - L.i.fits* e a ii ,• ■ C'r«*.r • d GO CM i-4 in in H o r- >- CO m 00 vO 00 at 'IZ-’Hity- x...... - t-* a,. V / . ' cn 0> 3 c m m c m Mt1 - H fy ,.V^'-H at t r c; (44 vj 4,^ r-l 5 O a vo o <;.• £ 62 44 a 44 r* ■<*’) r^ m 49 * 48 H }4 C/5 *"4 G 04 44 O H cn CM 32 18 O o o CM r4 CM 45 CM M4 o CO H H 60 td r4 H r4 CM. iisr4 . ©cpur Ivy <: 4J<0 mich&vy * 1 * 33 u V,-i *,. vc *’ ^ l»i*:s •• In’ *f Ktm' 13 > 01 escisser i-^r. 0) t9 u 44 at CM CM cn cn CM CM CM O •e m m m m in in « in H a^ Ov ov ov ov ov ov OV oo r-4 i-4 t-4 t-4 r-4 i-4 r-4 t-4 .fc' .tv "' '. .U4 d 44 d 'V : a) at d at - ' CL) a* O o CO a •n to S- CO co O *»4 C/5 4J •> ■ "V - t4 u U4 ,Q «1 • CD at O •*4 0) -r4 > > • / 4 V; •i4 • t-4 -i4 . . 4-i T) Pn .fc 44 44 - r-4 " V . ' . Kx . X ” ai ■ - oi at at 09 t4 •H a) cq r-4 .£ a M O. 4d Or ffi d O »4 d o d O d v y o 44 \ — (U o O >V o »r-l 44 44 44 44 > w d 03 d to 44 o E -r-4 *H •H *t4 •r4 CL) d at d a) d 0 Or Pu Or a. Pd t-4 at r-4 Ol r-4 •H *- v . 3 ‘ 0 a> at COCO CO CO •r4 43 t 4 JG -»4 O 3 r-4 O m B at S at B (U cr* (XJG rM -Id rid M 0) a. Or at & o M t4 U u B CM Pu O a. o rid w 3 c o o Q O 3 t-4 at m at oo r-4 < *“) V/ to co . p* >* OS Pd - 16 - - * - - r — about the four per cent level of probability (F = 2 .1 2 v F^ = 2.03, F0 .0i = 2.68). The odds are therefore about one in 25 that this difference is the result of chance. If the Jaimes River sample were eliminated from the calculations, the remaining samples would show no great departure from homogeneity. It is reasonable, on the basis of mean vertebral number, to conclude that fish sampled from the James River may belong to a different population than those sampled from other areas.of Chesapeake Bay and its tributaries. Because the statistical probability of this conclusion is < - + r not particularly strong, the question should be investigated further. Variation in Vertebral Counts with Length f ■el v** Sample 1 from York Spit was divided into two groups by standard length, 22 to 37 and 38 to 58 millimeters, to determine the relationship <*• v. •.tf ; Vi between numbers of vertebrae and size of fish. Mean numbers of vertebrae and sums of squares of deviations from the means were calculated for each group. Odds were about 50:50 that the difference of 0.05 vertebrae could occur by chance, thus, there is no evidence that non-random sampling of anchovy with respect to size of fish would introduce a bias into the t ■ * if t’ ’ , S f S a'- estimation of mean vertebral number. "C ' ... Oft £*• *4 r - •<' • . * * c A j t ‘"k «s* • y Precaudal and Caudal Vertebrae I i h C'lt K *5 Hi t/l — «...... Precaudal vertebral counts for the 704 individuals listed in tablel3 are given in table 14. Analysis of variance gave odds of less than one in 1 0 0 that the eight samples were drawn from a population homogeneous with respect to numbers of precaudal vertebrae (F = 2.94, Fq = 2.68). The variances of individual samples, as indicated by their coefficients -’ykset Z&^-XiMi^ .-,£%.7. CM 00 o rv rH r>. CO o •• • ••• •# C$1 H r-4 CM rH CM CM rH CM O MMu XH CO OS d 5! O 1 CM • o NO o o o CM NO o o o No rH rH r*i rH CO > © a; rH a a , cw o cn >{ CO u c m © XS Mf in ON NO rs. © ON u 00 00 NO 00 ON in CO © NO © •M ( • • «• ••«• ta *S CO cn m O CM CM © rH & V-f H tH CM «H H CM Cft k cn O \ IH CM t -i O G o v •- •‘■i- «H NO 00 Mf rH 00 o ON U ON o o O CO o o 00 CO © • ••• »■ • © « *H ON o o o ON © © ON .' .,:> r* J * i ■ •»' k I rH CM CM CM iH CM CM rH ’ *" CO c$ a © i ;}.■ i O CM • ' » ■■Cl, -■ '•■ -l i. " — £ © t c n © O 83 k <0 V CM © »M 3 3 H m © r M G k © n \ :ac?*v< I s i' •§ © © T J ** ' * ,3 “ X I P 3 5 B © oo •• • • rH « • t—4 O U P © j *■> Y- , . .. > -*4t ■: : ir i © © - - ' . •.. i k • . 4. ■ 'r" k rs. a p-< cn . iw CM o O CM CM © © CM CM CM CM in m m m m in m m os 83 ON ON ON 3 ON ON ON ON On k k rH rH rH rH rH rH rH r-4 © © JQ X3 • •• e e M •• • M 4J M • p CU o G r© a Or CU M S3 © © © © © © © O tw cn a *”3 cn cn cn o O (0 © n /TV a •U O O . o © - in CO •h © -: on/ r.h^ N^( . NH jj p OO k k XS s -*■ 4 > -•» © © •M L i ? • \ c > > •t-4 •H CO CS Cd . i- r*- - C V7SN cn u o O © o O >% ■M 4J M M > d d M o •rl •H •H •rl •rl G § d d d © a, a a a od © © •rl w. & i cn cn j d .d o 3 1 x: © © © tk O* o M ,« © a c« © © 4-5 k u M M S Or CM ^5 k X3 •t-4 O o O O © © © rH fc* M 8 >< P* >* >4 © 03 ed w of variation (C), also were heterogeneous (chi square = 72.16; P < 0.01), and it was interesting that the sample with lowest means (James River and '• - * I Elk Point) also had the largest variances. ‘ j There was some indication also of. heterogeneity in caudal *i vertebral counts (F = 2.33, F q = 2.03) and in their variances (chi square 32.39; P <0.01). Counts of caudal vertebrae (Table 15) were more variable * than total counts, and this may have been caused partially by failure to observe consistent criteria for separation of the two regions of the backbone, I ‘ r V .• * Tester (1937) found that the best method of separating the two parts of the i i i vertebral column was to recognize the first closed haemal arch as the first i ' caudal vertebra. This was not easy in this study because the lateral bridge ' ' " T could not always be seen clearly in side view. ; © co. ^ f Jo-*. to C M ?*»• ir» ir, * *• tft j; ' t ' * Structural Variations f ■ « * } & * In the four samples from York Spit, the James River and ’■ ; ! Rappahannock River 80 samples, each centrum bore a single neural arch and spine. The Rappahannock 50 samples contained four skeletons which had one extra neural and haemal spine each, and one skeleton lacked a pair of , ,*s» to - , * ' neural spines. One skeleton in the Elk Point sample bore an extra pair of neural spines, -■? 1 i * } ’** c.;'- . c* ft x , a % : 3* 4 ,t ..1 .• } , Aside from these variations in neural and haemal spines no structural abnormalities were observed among the 704 Anchoa.skeletons. Two individuals had a slight twist in the caudal section of the vertebral i column, but the number and size of vertebrae was not affected. ] o vO cn © o »n m rH O • •• •• «•• < 0 CM CM CM CO cn CM CM CO * h tt Dorsal, anal and pectoral fin rays were counted in the 704 specimens of Anchoa mitchilli. The count began with the most anterior ray, however small, and those branched at their distal ends were counted * as one, ^ • ! I h ' i ‘ I «• v vf. W - .*? -i ‘ - ‘*'i , i ■ . \ * w tit*. ' > | : }••! « O' C*' t* i . S . V, i frj - rt( »*n ^ --U ’ ’ 1 Dorsal Fin Rays | Vi) - » i v Frequency distributions of dorsal fin rays are given in table 16, Analysis of variance gave odds of less than one in 100 that the samples were drawn from a population homogeneous with respect to mean dorsal ray number (F as 5.07; P^O.Ol) , Inspection of"the~mean counts indicated that the Rappahannock River 50 and Elk Point samples were responsible for a consider- ■ i - i able amount of the variance between samples. Repeating the analysis with i ■■ *■ i 7- i- •• 4" ■? '■» '■» '■Tv these two samples eliminated gave no further indication that the samples *• ; i • could have been drawn from different populations. : Differences between mean dorsal ray counts for the Rappahannock 50 and 80 samples may be related to a salinity difference between the two areas. Differences in mean vertebral counts which are apparently related , ’i-i v ’C*'s :?■ i ,?*. to differences in salinity have been reported by Schmidt (1917, 1920), ijubbs (1925) and Blackburn (1950) . j ^ yv '*•» ■ I S.- r-;; .5' •• 0 ■ O ' :L * Anal Fin Rays •> I | .;'c 1 ? - ,• ; - - •. -* Analysis of variance gave odds of less than one in 100 that the I samples in table 17 could have been drawn from a homogeneous population ■ i with respect to numbers of anal fin rays. Lower counts predominated in I ' :: the lower part of the Bay, and high counts in the upper part, from the 00 CM vO co CO os OS 0 • . • • • •• • •• CO CO CO CO It to co If ... >4 .... § HI • 0 sO 0 0 0 'CM sO 0 u O 0 0 0 0 n» CO 0 •+\ (0 ss rl rl rl rl rl G ‘ %* O 1”' it •rJ - K> 44 ', 1. V «4 'i/ •rl .... - •-- A* > *-; & . !*v. CO -» 4* O CM 'r-s r- p» vO r r>. vO . VO m c if, u cn CN! co Ps P s oo » * • o < r o vO o o o CM M > o 45 o sf o o o p » oo o X r H • u CM a\ vo UN CT* OS vfr CM CM os OS I p * • P s • os cn UN CO o o * w cn f " s p . MI r>s oo cn o OS > CM cn PS. p*. P s cn os 05 *3. oj S3 •* cn rH c m cn cn CM CM cn .i-wzJ rn*> © ft U4 X > O .'8Vi- c 33 rH Ps. rH CN P s sCf- O so o cn rH cn c m CM rH rH **4 X cn •M 1 0 fc\*6 ;i © •p4 rt at X •H SX vO ps :ounts for the Rappahannock River 50 and 80 samples differ in a manner that say be related to a salinity difference. A series of samples would have o be examined from each area of the Rappahannock River in order to confirm his indication, especially in view of the reverse variation exhibited up ad down Chesapeake Bay. 1, sctoral Fin Rays Table 18 gives the frequency distributions of pectoral fin-ray unts. Analysis of variance gave odds of less than one in 100 that the raples could have been drawn from a population homogeneous for pectoral y number (F » 5.64; P < 0.01).The two extreme values were obtained from nples collected off the mouth of the York River, and in the absence of ler evidence it is not possible to explain the peculiar pattern of “iation. ationshlp Between Numbers of Body Segments and Fin Rays There was no apparent correlation in the anchovy between vertebral fin-ray counts (Table 19) • Usually there were two dorsal and two anal rays per myomere, but the fins occupied relatively small portions of il body length, only about eight and fourteen centra, respectively. xi other clupeoid fishes (Gwyn, 1940) fin rays are not formed for some after hatching, hence are affected by a different set of environmental itions than those that prevail during the embryonic period, when »br®l number is determined. * j Frequency Distributions of Pectoral Rays in Anchoa roltchilli, Means, Suras of Squares and Deviations from xs ss IM w «H •8 u <4-1 ;§ a ) v •H *H > u ■: 4 - . j ", . '. i , T ' i ' ’’ ' I" is ;**V- *r- ’••-J -*'OVA $ M* 5 »• ^ '* v<’ i? ; ' ft v ?Ut- i- e *\ 'n, 1 jw;? ? i'' /’Orpulnci. '• if«“. * ■ -. r ztifzi.is hr .?.%..*•■•.'?* -i*n; n c ^ ;g, '»y ?;*■ r-per:***- ?e> TABLE 19 '•; -• ■' . i t* : >-*'x ->*; i '•• .: i>. ..< ?r . ' ** ~ < ih- • , <.?« 16':. Correlation Coefficients and Levels of Probability for Several Meristic / ».-* . . 'J. ' .f . * : > r . - .• •» ■- x u ; v Characters Counted in the York Spit 1 Sample of Anchoa ciitchilli. l * .i. : * — :u. - a /• .*.»•* tv ; ' ,/A . r Correlation Characters Coefficient Probability • "t- ' > •• Total Vertebrae and Dorsal Rays 0*041 >0.05 > J* * .o'- * ^ ^»-V -v ,* ■ Total Vertebrae and Anal Rays 0.151 >0.05 * . -.-<«> i*» ti: *- * ^ L. X C C ■.. c. -0 :•. „ ji • . yi - • - f0.05 Caudal Vertebrae and Anal Rays 0o202 >0,05 . - . s.-. ~ ;ro -* ■ j. c. > ; >4 . •-••r ,';r ^ ;..e- J. i *. 'Z fUv Dorsal Rays and Anal Rays 0.119 >0.05. * ’.4 . ^ - * . * .‘c jyijet to iongtlii or .viie athev c- a meter -vsrias vUh. / \& *.■ t'rtar hah'd, it Las* be**. sJ* (Hildebrand« -943 s errs '' ..'kL~ *.*£ vc.; titbr&e decrease fxon nrri'n c** >, •• -ith •••etc .-.anr?? t. i /■ ** .■?* »i v> I % ’hi* ir a ?.? ,*p.±T'->Ztig$~ ~ Mttve, ur ; • f.he in??horn •habitat rr i&e *-•.«•* * 4 . ‘t ror.'f.. .ii.* . .-cv, ', f ) 92!. r 3-liicxbutta, 19K*} ^.oiciatic •.;Utirr.ct.ert» ' ’ * iivr.iv.} ^ ‘rb -%re vr-as evir n’^o .vi* tv ^orc^- c.riM t' oi ’ '** ; *>*.a o'. - 'd.X7\ rj.itrba -'.-v ■ f»n /•-* •’> t- COMPARISON OF MERISTIC VARIATION IN THE EEL AND ANCHOVY Number of Vertebrae 'V? *>•-» i ;There is no indication from statistical analysis of vertebral counts of Anguilla rostrata that samples from Nova Scotia, New Hampshire and Virginia were drawn from separate populations. Numbers of vertebrae V . y ’ * v *. : ■' _•. _ , are known to be affected by environmental factors, especially temperature, during early development (Jordan, 1892; Schmidt, 1921a; Hubbs, 1926; a:r m f j , '4' •••'• * ,• ;*•-■ - i:> ■■ J - «. •• •• ■ ’• *- *-* Tester, 1938; McHugh, 1951). Mean vertebral numbers tend to vary with ?t*» ' ’i < ■’>; • *- r - ‘C**-. V y .. "‘‘“'i - ■'*• • '- -4.- .i 4? . •«• 4 '* *■ latitude in species that spawn in coastal waters, populations in colder i ' j ’■*«,.. *, * „• > -■ .••<•*: 5» n ^ it* *-.■ *** , .uv.- northern areas tending to have higher mean numbers of vertebrae than those " £• t ^ * ' * *■ $ n 4. * ' ^ * ''' * ^ ' ' ■*• * v V i *. -. • *■ 'A £ * - i « KA *£ - ’V * •* found in warmer southern waters. The American eel, however, spawns in the :• 5, ^ . - :r ■* - *'• r~-;' *. ...• .-v. vicinity of the Sargasso Sea and migrates to coastal waters to spend most p n ;; -** .v -- #■ .».■ v ; y- • •«* .-.tV i of its life. It is probable that meristic characters are fixed before the *4? • 't.i i '• »**•■' '■ . ul > viT i ' -i-Z i * -4 , 1. ‘ ~..'a ... ■■ . .• ■« elvers reach the Atlantic coast, hence under rather uniform environmental ■ .. “ i.;i -*■ v V •' £.'■?■■■£.• e-r.P' .5?. '..'V .; !* v r. - , -'-i * conditions. Therefore, the heterogeneity demonstrated in Schmidt’s data (Table I) may have been caused by differential distribution up and down ^ 4 .v ?.u| '-i P 4 > * v ’jjl.;' ' i; J v ; ■ .*■ ‘ * . • ..V .» f .. 4 * U & l the coast with respect to length, or some other character that varies with 5-. * ’ - ' *• ' ^ '' r=. ^ f. % '■>*■ ’ X- >C ^ <• *" * Sr.* , > number of vertebrae. 1 > j ~:. > ■- «'h U > ^ r s t . the other hand, it has been shown (Hildebrand, 1943) that in Anctjoa mltchiili mean numbers of vertebrae decrease from north to south along the Atlantic coast. Presumably this is a temperature-controlled phenomenon related to the inshore habitat of the species .j In some f ishes (Schmidt 1917, 1920; Hubbs, 1925; Blackburn, 1950) meristic characters vary directly with salinity. There were some evidence of heterogeneity of vertebral counts within Chesapeake Bay, but the variation showed no regular pattern with salinity or any other known environmental factor. - 2 1 - Relative variability of vertebral counts, as expressed by the coefficient of variation, was somewhat greater in jthe anchovy. This might be expected in view of the greater variability in environmental conditions in the inshore zone where anchovies undergo their embryonic development. r ' >•. • • Variation in Vertebral Counts with Length A significant relationship exists between numbers of vertebrae length in the eel (Table 4), but there is no indication of a similar relationship in the anchovy. Usually, when such a relationship is demonstrated, it is inferred that size is proportional to age, hence the smaller fish developed under different sets of environmental conditions than the larger fish. It is possible that the differences between Schmidt*s mean vertebral counts were caused by biased sampling with respect to length for it is known that in estuarine areas, young fish often are distributed very precisely up and down stream with respect to size (Haven, 1957). |A^riations in body length also could be at least partially related to variations in numbers of vertebrae, for it seems reasonable to suppose that in two fish of equal age and growth rate, the one with the greatest number of vertebrae would also be the longest.sKu, '.-ru < •->c : Unity tic * Structural Variations Variations in structure of the centrum and its processes, in the form of extra spines or evidence of fusion of adjacent centra, were observed in about 59 per cent of eels, but in only about 1 per cent of anchovies. A significant difference in mean vertebral number was found between "normal" and "abnormal” eels, the former averaging about one-half vertebra less. - 2 2 - However, ratios of “abnormal" to “normal" backbones appeared to be constant * regardless of locality, hence the influence of “abnormalities" upon mean vertebral counts need not be considered in measuring latitudinal variation, as bng as counts derived from such fish are made according to pre- established criteria. ; r #ou‘l : i •• . ..4 = a* • I f ■ '--i ."** rr., Number of Fin Rays ;• ; ?F- (*■?.> t ' " - •* Analysis of the eel sample means for dorsal, anal and total fin- v •• - ‘ '• • i •' " - • • ray counts gave no indication that the three samples could have been drawn from separate populations. On the other hand, mean dorsal, anal and pectoral fin-ray counts in Anchoa mitchilli all gave evidence of heterogeneity, although the samples responsible were not the same for all fins. Variations in fin-ray numbers may be caused by environmental factors and since the various fins complete their development at different times their individual patterns of variability need not be similar. Little is known concerning the extent of anchovy migrations within Chesapeake Bay. However, if they migrate extensively, one would not expect to find variations that could be related to latitude, temperature or salinity. : ■- v r-v . v, t .There was no particular evidence that heterogeneity in Anchoa fin-ray counts was associated with salinity variations, as the work of Schmidt (1917, 1920), Hubbs (1925), and Blackburn (1950) might suggest, except in the Rappahannock River where a sample taken 80 miles above the mouth, in fresh water, had lower dorsal and anal counts than one taken 30 Kilos downstream in brackish water. -.-..I r -v ; ' t ' * ®ay be that observed differences in Anchoa fin-ray counts wet© due to variation in environmental factors during the period when fin- ray numbers were determined, factors perhaps more variable than during the smbryo stage when vertebral number was determined. It also seems possible '.hat these differences could arise from sampling a statistically complex biological population and, if so, a great many more samples, particularly rom year to year in the same locations, would need to be examined to define he underlying patterns more clearly. - ^ * *• ! - vOFin-ray counts were relatively much more variable than counts of artebrae. In the eel, counts of total rays were somewhat less variable lan counts for individual; fins, and this probably reflects the difficulty : distinguishing between caudal, dorsal and anal rays inaa fin that is ^ y.-\ mtinuous. Dorsal fin-ray counts in the anchovy were relatively less J‘„ : l ' \\t ■ vatfacie u Ck. . ' -■ * ...* ;■ ' • ’ ; • rtable than in the eel, but variability of anal rays was quite similar . ,u-. ,• New snite • ‘i'ui ^ j , 's, - both species. t-.j *<«. Ch&9&p&$Kt* .£sy -sr?d % . r i / ' • f. i-tv c-. ■> iationship Between Numbers of Body Segments and Fin Rays ' v '■"- v -Variations in numbers of dorsal and anal fin rays were strongly sociated with numbers of vertebrae in the eel, but, in the anchovy, no .h relationship was evident. In the eel these fins, especially the :w sal, occupy linear distances similar to that occupied by the vertebral umn. The primitive median fin, though present in larval anchovies, does persist in adults and the separated remnants that form the adult fins associated with a relatively small fraction of the total number of oeres whereas the vertebral column is immediately associated with the r segments. Such differences in physical association of vertebrae and rays with body segments could be largely responsible for the different Is of correlation between these characters in the two species. - 2 4 - Another clue to the lack of relationship between vertebral and fin-ray numbers in the anchovy is contained in the work of Ford and Bull (1926) on Clupea harengus and McHugh (1942, 1951) on Clupea pallasii and Engraulis mordax respectively. Following development of dorsal and anal fin rays in these clupeoids, and while they are still developing, these fins migrate forward over a. considerable number of myomeres to take up their adult positions. It. is obvious that there cannot be a very close association between body segments and fin rays under such circumstances, and that theie is no reason to expect that a relationship should exist. 1 \ £ f.:»• numbtirs •©£ • -tLn r&/a- \*&em l y SUMMARY V > -1 *v. s -V: , ‘ ■ >■»**] -•* rlo-^h i ;>-o A study of variation of meristic characters of Anguilla rostrata ! »..•,« • i •••* ‘3 *.** f- --'-V ' ’ *•:, .t' f,r .t I ••••■' . rom Nova Scotia, New Hampshire and Virginia, and Anchoa mitchilli from .2 . '• •* > ■ by- ’."‘vi'C,51 -***>-;■ i«».»:i . r -• ive locations within Chesapeake Bay and its tributaries was made to compare ariabillty in numbers of vertebrae and fin rays in two fish species that • ;\* '.vi. •• ‘ L\&z er.cusi nr-let p i n the 1. •; Iffer widely in numbers of body segments. . *• r„ *. c . • ' ' v? ■.■*. •£* s,• •• duri'CiT T 1' ' £ f r*7 J--, . f "i ?§ Statistical analysis of numbers of vertebrae and fin rays in eel ■’ « md anchovy suggest that populations along the Atlantic coast are hetero- neous. Whether these differences are genetic or environmental in origin, whether they are caused by segregation according to other factors, it not possible to determine. This has no particular bearing on the major erpose of this study which was to compare variation between the two species Relative variability of vertebral counts, as expressed by the efficient of variation (C), was somewhat greater in the anchovy. This may due to difference in environmental factors during embryonic development the two species. Fin-ray counts were relatively more variable than nts ©£ vertebrae in both species. - 25 » Numbers of vertebrae are significantly related to body length in the eel, but in the anchovy there is no evidence that variations in these characters are related. ’ ' Structural variations occurred in about 59 per cent of eels, but in only one per cent of anchovies. "Abnormal" eel skeletons averaged about one-half vertebra less than "normal*1 skeletons, but as the ratios of "abnormal** to "normal" backbones appeared to be constant regardless of locality, the effect of "abnormalities’* on mean vertebral counts was of no consequence. >In Anguilla, variations in numbers of fin rays were strongly related to numbers of vertebrae, but no relationship was evident between these characters in Anchoa. It is suggested that these differences may be related to the degree of physical association of vertebral centra and fin rays to body segments. It is also possible that the poor relationship ‘in the anchovy may be due to differential growth in the post larval stage, as has been shown for other clupeoids, during which time fin migrations occur. " • *• - 26 - t t t f .RAT UR E CITED Anderson, A. W. and E. A, Power M *. t4’ / *•- 1 -m***: - ' * 1955* Fishery Statistics of the United States. U. S. Fish and ■ ■■?**.- Wildlife Serv., Statistical Digest 41. 446 pp. Blackburn, M. • ~ " 1950. A biological study of the anchovy, Engraulis australis (White), in Australian waters. Australian Jour. Marine and Freshwater i * . Res. 1(1): 3-84. Clark, Frances N. 1936. Variations in the number of vertebrae of the sardine, Sardinops . n caerulea (Girard). Cope la 1936(3): 147-150. Ford, E. . 0. :• . rv *•' 1933. The number of vertebrae in the herring and its variations. Jour. Conseil 8(2): 211-222. '• - ;r Ford, E. and H. 0. Bull •' - - ^ ^ - 1926. Abnormal vertebrae in herrings. Jour. Marine Biol. Assoc. 14: 599-517. . ‘‘.if.-ii.: :*•». * - w * ‘* Gabriel, M. L.. ’-1-' **■* *-v v-iv-.; to -*v:• 1944. Factors affecting the number and form of vertebrae in Fundulus heteroclitus• Jour. Exp. Zool• 95: 105-143. Gwyn, Agnes M. ^. j ; v «*. - : - -A 1940. The development of the vertebral column of the Pacific herring, Clupea pallasii. Jour. Fish. Res. Bd. Canada 5(1): 11-22. Hart, John L. i .v-: ic* *. v ist. * ;< ' 1933. Statistical studies on the British Columbia pilchard: Vertebrae counts. Trans. Roy. Soc. Canada 1933(3), P. 27£ (5): 79-85. - 27 - Haven, D. 1957. Distribution, growth and availability of juvenile croaker, Micropogon undulatus. in Virginia. Ecology 38(1): 88-97. Heincke, Friedrich ?. v. :.a n 1898. Naturgeschichte des Herings: I - Die Lokalforraan und die Wanderungen des Herings in den europaischen Merren. Berlin: i ,223 pp. Hildebrand, S. F. vtr- in I - i 1943. A review of the American anchovies (Family Engraulidae). Bull. Bing. Oceanogr. Coll. 8(2). 165 pp. Hildebrand, S. F. and W. C, Schroeder ( .11. i-v*.. ^ 1927. Fishes of Chesapeake Bay. Bull. U. S. Bur. Fish. 43. 388 pp. Hubbs, Carl L. 1925. Racial and seasonal variation in the Pacific Herring, California sardine and California anchovy. California Fish and Game Comm., Fish. Bull 8. 23 pp. ° ; * 1926. The structural consequences of modifications of the development ■'A. rate in fishes, considered in reference to certain problems of evolution. Amer. Nat. 60: 57-81. _ Hubbs, C. L. and A. Perlmutter , t. * cions* in >v. ’ 1942. Biometric comparison of several samples, with particular reference to racial investigations. Amer. Nat. 76: 582-592. Jordan, D. S. • c-.*-.*•, t. «*:.«'ix I \ „ :x 1892. Relations of temperature to vertebrae among fishes. Proc. U. S. Nat. Hus. 14: 107-120. Kandler, R. r • • **-. "* 1932, Uncertainty in the determination of vertebral number as a i is ■ result of apparent deformities. Jour. Gonseil 7(3): 373-385. Masstnann, W. H., E. C. Ladd and H. N. McCutcheon 1952. A surface trawl for sampling young fishes in tidal rivers. . vi ; . No. Amer. Wildlife Conf. 17: 386-392. McHugh, J. L. ^ mf-ri.3tLs c- - 1942. Variation of vertebral centra in young Pacific herring (Clupea pallasii). Jour. Fish. Res. Bd. Canada 5(4): 347-360. 1951. Meristic variations and populations of northern anchovy (Engraulis mordax mordax)♦ Bull. Scripps Inst. Oceanogr. -6(3): 123-160. Rounsefell, G. A. • «£' terrla# .-V 1930. Contribution to the biology of Pacific herring. Clupea pallasii. and the condition of the fishery in Alaska. Bull. U. S. Bur. Fish. 45: 227-330. - - Cuts,!CU-.C Schmidt, Johannes 1914. First report on eel investigations. Rapports et Process- Verbaux des Reunions, Conseil Perm. Int. Exp 1. Her. 18. 1916. Second report on eel investigations. Ibid. 23. 1917. Racial investigations. I. Zoarces viviparu9 L. and local races of the same. Compt. Rend. Lab. Carlsberg 13(3): 279-397. 1920. Racial investigations. V. Experimental investigations with Zoarces viviparus L. Ibid. 14(9): 1-14. 1921a. Racial investigations. VIX. Annual fluctuations of racial characters in Zoarces viviparus L. Ibid. 14(15); 19-23. Snedecor, G. W. - 1946. Statistical Methods. 4th Edit. Collegiate Press, Ames, Iowa, 485 pp. Taning, A. Vedel 1944. Experiments on meristic and other characters in fishes. I. On the influence of temperature on some meristic characters in sea trout and the fixation-period of these characters. Medd. fra. Koran. Danmarks Fisk. Havunder., Ser. Fish. 11(3): 66 pp. Tester, A. L. 1937. Populations of herring (Clupea pallasii) in the coastal waters of British Columbia. Jour. Biol. Bd. Canada 3: 108-144, 1938. Variation in the mean vertebral count of herring (Clupea pallasii) with water temperature. Jour. Conseil 13(1): 71-75.» H rH rH u 00 * oo '"yss-; o co d J o St OS n* Os o in H r CM rH h • . oo CM >s O r-4 r-4 3 CM Cd •3 3 t—4 3 CO m sO r-4 3 r-4 r-4, 3 •\ CM M 3 O 3 f t 3 00 r-4 CO CM O - *4-1 5 . CM O O 3 3 >» to Q* 3 CO • H, .*■. 3 CO a ; O O CM -r-4 *W 3 U O > P t-4 X 3 &y *i-4 U i U 3 *4-1 i 4J X O *0 § 3 •ft P 3 U f t 2 ! r-4 3 •H 3 r* •H XS >1 *N > U 4J 3 O 3 M •r-4 O a 3 3 3 3 r-4 o ♦ri 3 3 4J 3 CO 3 3 P 2 3 a 35 •t-4 cr M o 3 60 3 3 3 > Sc M M X So o 3 •H 5* 4J x 2 2 > O 00 o • ♦ us ! 4! . 5 • 5n o NO o vO B (X * O 3 css GN us O U O ON fs. US »w M t X • • • O t tfV ON CM n X o CM r-4 d CO vO r*. US o ft) C/3 •N •s ■H •H 00 ON r- 4~) O to d 0) > 3 ft) cr CM US 00 O ft) d 00 VO 00 M d • « 9 c m Cn