ARTICLE

Movement patterns and dispersal potential of Pecos bluntnoseshiner (Notropis simus pecosensis) revealed using otolithmicrochemistryNathan M. Chase, Colleen A. Caldwell, Scott A. Carleton, William R. Gould, and James A. Hobbs

Abstract: Natal origin and dispersal potential of the federally threatened Pecos bluntnose shiner (Notropis simus pecosensis) weresuccessfully characterized using otolith microchemistry and swimming performance trials. Strontium isotope ratios (87Sr:86Sr)of otoliths within the resident plains killifish (Fundulus zebrinus) were successfully used as a surrogate for strontium isotope ratiosin water and revealed three isotopically distinct reaches throughout 297 km of the Pecos River, New Mexico, USA. Two differentlife history movement patterns were revealed in Pecos bluntnose shiner. Eggs and fry were either retained in upper river reachesor passively dispersed downriver followed by upriver movement during the first year of life, with some fish achieving aminimum movement of 56 km. Swimming ability of Pecos bluntnose shiner confirmed upper critical swimming speeds (Ucrit) ashigh as 43.8 cm·s−1 and 20.6 body lengths·s−1 in 30 days posthatch fish. Strong swimming ability early in life supports ourobservations of upriver movement using otolith microchemistry and confirms movement patterns that were previously un-known for the species. Understanding patterns of dispersal of this and other small-bodied fishes using otolith microchemistrymay help redirect conservation and management efforts for Great Plains fishes.

Résumé : L'origine natale et le potentiel de dispersion du méné (Notropis simus pecosensis), une espèce menacée au niveau fédéral,ont été caractérisés avec succès par la microchimie des otolithes et des essais de performance natatoire. Les rapports d'isotopesde strontium (87Sr:86Sr) des otolithes chez le fondule résident (Fundulus zebrinus) ont été utilisés comme substituts des rapportsd'isotopes de strontium dans l'eau et ont révélé trois tronçons distincts sur le plan isotopique le long de 297 km de la rivière Pecos(Nouveau-Mexique, États-Unis). Deux motifs distincts de déplacements associés au cycle biologique ont été décelés chez méné.Les œufs et les alevins étaient soit retenus dans des tronçons supérieurs de la rivière ou dispersés passivement vers l'aval pourensuite se déplacer vers l'amont durant la première année de vie, le déplacement minimum de certains poissons atteignant56 km. La capacité natatoire de méné a confirmé des vitesses de nage critiques supérieures (Ucrit) pouvant atteindre 43,8 cm·s–1

et 20,6 longueurs du corps·s–1 30 jours après l'éclosion des poissons. Une bonne capacité natatoire tôt durant la vie appuie nosobservations de déplacements vers l'amont obtenues de la microchimie des otolithes et confirme des motifs de déplacementjusqu'ici inconnus pour cette espèce. La compréhension des motifs de dispersion de cette espèce et d'autres poissons a petit corpsa l'aide de la microchimie des otolithes pourrait aider a réorienter les efforts de conservation et de gestion des poissons desgrandes plaines de l'Ouest. [Traduit par la Rédaction]

IntroductionRivers throughout the southwestern Great Plains have experi-

enced dramatic changes in flow regimes over the last 100 years.These changes are the result of damming; channelization; waterdiversion; groundwater pumping for municipal, industrial, andagricultural uses; and frequent and persistent droughts (Doddset al. 2004; Hoagstrom et al. 2008b; Durham and Wilde 2009).The results of flow intermittency and decreased habitat complex-ity have negatively affected pelagic broadcast spawning fishes(Bestgen et al. 1989; Durham and Wilde 2009; Hoagstrom et al.2011). For these fishes, spawning occurs throughout the summerand is cued to not only high-flow events (synchrony), but across arange of flow conditions. This spreads reproduction throughout aseason, resulting in multiple opportunities (asynchrony) as a bet-hedging strategy to increase chances of young surviving to recruit-ment (Durham and Wilde 2005; Durham and Wilde 2008).

Broadcast spawning fishes utilize a reproductive strategy inwhich the female releases ova into the water column whereuponthe male(s) fertilize them. The eggs are nonadhesive and semi-buoyant, and as long as water velocity is sufficient, the propaguleswill remain suspended within the water column and drift down-river while development occurs (Platania and Altenbach 1998;Cowley et al. 2009). Retention of eggs and fry in slack-waternursery habitat reduces the chances of continued downriverdisplacement and may be key to successful recruitment into thepopulation (Dudley and Platania 2007; Hoagstrom and Turner2015). In contrast, eggs and fry are at risk if swept into unfavorablehabitat such as irrigation canals or into large impoundmentswhere they would perish (Dudley and Platania 2007). Thus, under-standing life history movement patterns of broadcast spawningfishes in response to both natural and anthropogenic factors mayassist in management and recovery efforts for these fishes.

Received 29 December 2014. Accepted 19 May 2015.

Paper handled by Associate Editor Aaron Fisk.

N.M. Chase. New Mexico State University, Department of Fish, Wildlife and Conservation Ecology, Las Cruces, NM 88003, USA.C.A. Caldwell and S.A. Carleton. US Geological Survey, New Mexico Cooperative Fish and Wildlife Research Unit, Las Cruces, NM 88003, USA.W.R. Gould. New Mexico State University, Economics, Applied Statistics & International Business Department, Las Cruces, NM 88003, USA.J.A. Hobbs. University of California-Davis, Wildlife, Fish & Conservation Biology Department, Davis, CA 95616, USA.Corresponding author: Colleen A. Caldwell (e-mail: ccaldwel@nmsu.edu).

1575

Can. J. Fish. Aquat. Sci. 72: 1575–1583 (2015) dx.doi.org/10.1139/cjfas-2014-0574 Published at www.nrcresearchpress.com/cjfas on 22 June 2015.

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y D

IRE

CT

OR

AT

E O

F C

OL

DW

AT

ER

FIS

HE

RIE

S R

ES

on 1

0/08

/15

For

pers

onal

use

onl

y.

In the Pecos River, New Mexico, channel scouring and alteredflow regimes from water diversions, channelization, and largeimpoundments threaten a guild of broadcast spawning fishes thatinclude native speckled chub (Macrhybopsis aestivalis), Rio Grandeshiner (Notropis jemezanus), and Pecos bluntnose shiner (Notropissimus pecosensis; Platania and Altenbach 1998; Dudley and Platania2007; Hoagstrom et al. 2011). Of conservation interest is the feder-ally threatened Pecos bluntnose shiner. Once found throughout631 km of the Pecos River, the fish is now restricted to only 330 km(Platania 1995; Hoagstrom 2003a; Hoagstrom et al. 2008a). Thespecies was state-listed by New Mexico as endangered in 1975(NMDGF 2012) and federally listed as threatened in 1987 (USFWS1987). The fish is relatively short-lived, with a lifespan of 2 to3 years in the wild (Hatch et al. 1985; Bestgen and Platania 1990;Hoagstrom et al. 2008b). From May through the end of September,spawning was historically cued by high-flow events from snow-melt runoff and summer monsoon rains that increased flow froma few hours to a few days. Spawning cues currently include sum-mer precipitation, large volume of long duration water releasefrom a large impoundment (Sumner Dam), and flooding fromnonregulated tributary inputs (Hatch et al. 1985; Hoagstrom et al.2008b).

Persistence of Pecos bluntnose shiner is dependent upon thedispersal and hatch of eggs and whether larvae find refuge innursery habitat. Without an understanding of dispersal patternsor movement related to reproduction and eventual recruitmentof this fish, biologists can only presume environmental variablesthat limit successful reproduction and recruitment. Unfortu-nately, little is known of passive dispersal of egg and larvae as wellas movement patterns of this species (Platania and Altenbach1998; Hoagstrom et al. 2008a, 2008b). Movement of juvenile andadult Pecos bluntnose shiner have not been studied in depth;however, some notable size reduction downriver has been docu-mented (Hoagstrom et al. 2008b). The absence of adult Pecosbluntnose shiner in the southern-most occupied portion of theriver above Brantley Reservoir suggests that fish in this lower areaare either not recruiting into the population or are not survivingto maturity (Hoagstrom et al. 2008b). In addition, source areaswhere propagules deposit and larvae develop to eventually recruitinto the core population are unknown (Platania and Altenbach1998; Hoagstrom et al. 2008a). Ultimately, recruitment of Pecosbluntnose shiner is dependent upon where eggs disperse andhatch and if larvae find refuge. If the source of recruitment comesfrom downriver, those individuals likely must swim upriver tosuccessfully reproduce or their eggs potentially disperse furtherdownriver into Brantley Reservoir (Platania and Altenbach 1998;Dudley and Platania 2007). The success of this fish is further com-plicated by the timing and magnitude of water release from Sum-ner Dam to reduce evaporative loss between reservoirs andefficient water delivery for agricultural use. Without knowingdispersal patterns or movement related to reproduction of thespecies, managers could only presume environmental variablesthat affect successful reproduction and recruitment.

Future management efforts targeted at improving and facilitat-ing life history movement patterns of Pecos bluntnose shiner mayaid in its recovery. However, little is known about movementpatterns and dispersal potential of this species after larval devel-opment. The objectives of this research were to (i) assess the di-rection and extent of potential dispersal of Pecos bluntnose shinerby reconstructing natal origin using otolith microchemistry ofstrontium isotopes (87Sr:86Sr) from hatch until time of captureand (ii) perform controlled swimming performance trials to betterunderstand age-related dispersal potential in this species. Under-standing life history movement patterns in Great Plains fishes isan essential step in developing conservation strategies and guid-ing informed management actions to ensure persistence of thisunique guild of broadcast spawning fishes.

Materials and methods

Study areaCurrently, the Pecos bluntnose shiner is restricted to the Pecos

River main stem from Sumner Dam to Brantley Reservoir, a dis-tance of 330 km within three distinct areas (Tailwater, Rangelands,and Farmlands; Fig. 1; Hoagstrom 2003a, 2003b). The Tailwaterarea is the most northern portion of the river between SumnerDam and the confluence of the Pecos River at Taiban Creek andextends 33 km (Fig. 1). Release of sediment-free water from Sum-ner Dam leads to channel scour creating unsuitable habitat wherethe species has not been collected since 1999 (Hoagstrom 2003a;Hoagstrom et al. 2008b; Davenport 2010). The middle section(Rangelands) is characterized as having the most suitable fish hab-itat of shifting sand-bed and a braided river channel extendingfrom Taiban Creek to the Rio Hondo confluence for a total of155 km (Fig. 1). All size classes of Pecos bluntnose shiner have beenroutinely documented within this area of the river (Hoagstrom2003a, 2003b). The southernmost section (Farmlands) extendsfrom the Rio Hondo confluence to Brantley Reservoir for a total of142 km (Fig. 1) and is characterized as a deeply incised narrowchannel with a compacted river bed, modified for more effectivewater delivery (Tashjian 1993). Salinity is elevated in this lowerarea owing to the cumulative effects of diminished stream flow,increased evapotranspiration, saline irrigation return flows, andbrine aquifer intrusion (Hoagstrom et al. 2008a; Hoagstrom 2009).Adult Pecos bluntnose shiner are generally absent from this mostsouthern area (Hoagstrom et al. 2008b).

Bedrock and Pecos River water chemistryA geologic map of the study area revealed unique differences in

bedrock formation throughout the Pecos River drainage (NMBGR2003). The dominant bedrock throughout the northern section ofthe river is reflected by the Guadalupian Formation from thePermian period (270–260 million years ago). In contrast, the dom-inant bedrock throughout lower reaches of the river is reflectedby Piedmont alluvial slopes from the Holocene to early Pleisto-cene, which spans the most recent glaciations from 2.5 millionyears ago to present (NMBGR 2003; Walker and Geissman 2009).The Quaternary alluvium deposit begins about 16 km north ofRoswell (located near the monitoring site at Highway 70; Fig. 1)and extends through the Roswell Basin (southern portion of thefish’s range) (http://pubs.usgs.gov/ha/ha730/ch_c/C-text7.html, ac-cessed 19 May 2013). These formations consist of carbonate (lime-stone and dolomite) and evaporite (gypsum and halite). Thus,older bedrock in the northern portion of the river is reflected byhigher 87Sr:86Sr values, while younger bedrock in the southern por-tion of the river is reflected by lower 87Sr:86Sr values (Barnett-Johnsonet al. 2008; Hegg et al. 2013).

A total of four water samples were collected 23–25 April 2012during base flows as an initial assessment of 87Sr:86Sr values. Thefour sample locations were predecisional in an attempt to locatedifferences in 87Sr:86Sr reflecting dominant bedrock. Thus, twowater samples were collected from the most northern samplingsites within the Rangelands area at Willow and Highway 70, andtwo water samples were collected in the most southern samplingsites within the Farmlands area at Dexter and Highway 82 (Fig. 1).The water samples were analyzed using inductively coupledplasma mass spectrometry (ICPMS) at the University of California-Davis Interdisciplinary Center for Plasma Mass Spectrometry.

Fish collection and otolith preparationThe plains killifish (Fundulus zebrinus) is present throughout the

home range of the Pecos bluntnose shiner and rarely exhibitslarge-scale patterns of movement (Minckley and Klaassen 1969).We chose to use otolith 87Sr:86Sr values from the plains killifish asour reference (or surrogate) for 87Sr:86Sr in water as a more cost-efficient alternative to characterize 87Sr:86Sr values throughout

1576 Can. J. Fish. Aquat. Sci. Vol. 72, 2015

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y D

IRE

CT

OR

AT

E O

F C

OL

DW

AT

ER

FIS

HE

RIE

S R

ES

on 1

0/08

/15

For

pers

onal

use

onl

y.

the range of the Pecos bluntnose shiner. As a resident, the plainskillifish would presumably capture ambient water chemistry at aparticular location (from time of hatch to time of capture). Thus,a range of sizes of plains killifish (n = 97, range 19.4–55.4 mmstandard length, SL) were collected 7–9 November 2012 using a3.0 m × 1.2 m seine with 3.2 mm mesh. From north to south, thesample collection sites were Willow, 6 Mile Draw, CrockettDraw, Bosque, Gasline, Highway 70, Dexter, Lake Arthur Falls,and Highway 82 (Fig. 1).

Pecos bluntnose shiner were sampled on the same sample datesand within the same sample sites as the plains killifish. A range ofsizes of Pecos bluntnose shiner (n = 119, range 29.7–60.1 mm SL)

were also collected. From north to south, the collection sites wereWillow, 6 Mile Draw, Crockett Draw, Bosque, Gasline, High-way 70, and Dexter (Fig. 1). Pecos bluntnose shiner were absentfrom sample collections at Lake Arthur Falls and Highway 82(Fig. 1). Fish were collected just before the onset of winter to en-sure sufficient growth of young-of-year fish for capture in seinesand to detect movement using otoliths. Captured fish were eutha-nized, placed on dry ice, and transported to the laboratory. Sagit-tal otoliths were removed, placed into 1.5 mL vials with ultrapure(milli-Q) water, and cleaned using an ultrasonic, ultrapure waterbath for 5 min to remove organic tissue. Otoliths were then rinsedagain with ultrapure water, placed in new 1.5 mL vials, and al-

Fig. 1. Study area of the Pecos River and the sample collection sites. From north to south, the river is categorized according to anthropogenicinfluences of Sumner Dam (Tailwater), undisturbed open range (Rangelands), and agricultural influences (Farmlands).

Chase et al. 1577

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y D

IRE

CT

OR

AT

E O

F C

OL

DW

AT

ER

FIS

HE

RIE

S R

ES

on 1

0/08

/15

For

pers

onal

use

onl

y.

lowed to dry under a class 100 laminar-flow hood. After 48 h drytime, otoliths were mounted sulcus side up, affixed to a micro-scope slide with Crystalbond (Crystalbond 509, Ted Pella Inc., Red-ding, California), and sanded using a MTI Corporation UNIPOL1210 grinding–polishing machine (1200 grit sand paper wettedwith ultrapure water) to reveal the core to the edge (Thorrold et al.1998; Hobbs et al. 2010).

Laser ablation multicollector inductively coupled mass spec-trometry (LA-ICPMS) was used to assess 87Sr:86Sr in otoliths.Analysis was conducted at the University of California-Davis Inter-disciplinary Center for Plasma Mass Spectrometry using a NewWave Research UP213 laser ablation system coupled with a NuPlasma HR (Nu032) multiple-collection, high-resolution, double-focusing plasma mass spectrometer system. Line scans across theface of the otolith from the core to the edge generated 87Sr:86Srprofiles throughout the fish’s life. A scanning speed of 10 �m·s−1,beam width of 40 �m, laser pulse frequency of 10 Hz, and 65%laser power were used. Values for 87Sr:86Sr were normalized inrelation to 87Sr:88Sr (0.1135) to correct for instrumental mass frac-tionation. Instrumental accuracy was ensured using a modernmarine coral (an in-house calcium carbonate standard).

AgingAfter isotopic analysis, otoliths were photographed using a

Leica DME microscope with Leica ICC50 Camera Module and LeicaLive Image Building Software (LAS Software Version 4.4.0, Heer-brugg, Switzerland) to generate whole otolith photographs. Datafrom the isotopic analysis were overlaid following the ablationpath for each otolith, and the ages at which isotopic shifts oc-curred were then recorded for each fish. Otolith photographswere viewed by two independent analysts, and assigned ages werecompared (Miller and Storck 1984). Briefly stated, where age dis-crepancy occurred at greater than 10%, a third analysis was per-formed and ages were accepted if the third age analysis fell within10% of one of the first two aging attempts. If consensus could notbe reached, age for that fish was excluded. Of the total Pecosbluntnose shiner collected (n = 119), five fish could not be aged andwere excluded from the age classification.

Age at movementFor each Pecos bluntnose shiner, the age at which an isotopic

shift occurred from laser ablation analysis was identified fromdigital images using ImageJ software (version 1.48i, National Instituteof Health). Distance (�m) was calibrated using known length laserablation line recorded during 87Sr:86Sr analysis. Fish growth var-ies between the warmer summer months (majority of growthoccurring during this season) and cooler autumn–winter (growthis slow). Daily rings were easily counted the first year of life. Thus,age at which fish moved was binned into groups, with 0+earlyrepresenting within 30 days posthatch, 0+mid representing 30–60 days posthatch, 0+late representing from 60 days posthatch topre-annulus formation, 0+winter representing within the first an-nulus (winter), 1+early representing early growth after winter (sec-ond growing season), 1+mid representing midsummer growth(second growing season), and 1+late representing late summer–fall growth (second growing season).

Swimming performanceCaptive propagated Pecos bluntnose shiner were tested at

30 days (n = 30, mean 20.63 mm total length, TL), 60 days (n = 30,mean 33.93 mm TL), 90 days (n = 15, mean 46.33 mm TL) posthatchand wild-caught adults (n = 30, 69.13 mm TL) from the Pecos Riverusing a swim tunnel (Loligo Systems, Denmark). Fish youngerthan 30 days posthatch could not be tested with the staminatunnel (several 30-day fish escaped from the stamina tunnel re-sulting in test termination and were not included in calculations).Water quality was monitored and maintained such that it did notinfluence swimming performance among age classes. Fish were

acclimated at 1.0 cm·s−1 flow 1 week prior to the swimming trials.On test day, individual fish were placed in the stamina tunnel andallowed to acclimate for 1 h at 5.0 cm·s−1. Flow was increased by10.0 cm·s−1 increments at 5 min intervals until the fish fatiguedand became pinned against the back screen for more than 5 s(conclusion of the test). At the termination of each test, fish weremeasured for total and standard lengths (mm) and placed into arecovery tank. Critical swimming speed (Ucrit) was calculated us-ing the equation from Beamish (1978):

(1) Ucrit � Ui � ��ti · tii�1� × Uii�

and body lengths per second

(2) BL ·s�1 � Ucrit ·TL

where Ui = the full interval swam at the highest velocity (cm·s−1),Uii = the velocity increment (cm·s−1), ti = time (s) fish swam in thefinal increment until becoming pinned, tii = duration of each in-crement, and TL = total length of individual fish.

Data analysisFor both species, a ten-point moving average was used to

smooth the 87Sr:86Sr values for visual inspection. The fish wasdeemed a resident if no isotopic shifts were evident through vi-sual inspection of the full data profiles matching one locationthroughout the fish’s life and by comparing means betweenhalves of the line scans for each fish via t tests (JMP, version 10, SASInstitute Inc.). When 87Sr:86Sr values revealed an isotopic shift(i.e., clear breakpoint in the data indicating different reaches),then it was presumed the fish moved between areas of unique87Sr:86Sr chemistry. For example, an upriver movement was iden-tified if isotopic values near the otolith core were classified todownriver reaches and isotopic values near the otolith edge wereclassified to upriver reaches, or vice versa. When an isotopic shiftwas evident, isotopic values representing each location were par-titioned separately because fish were presumed to have spenttime within an isotopically distinct reach. We verified break-points in the 87Sr:86Sr values for each fish using a Student’s t testto determine if pre- and postbreakpoints were significantly differ-ent. For fish exhibiting two or more breakpoints, we used ananalysis of variance (ANOVA) to test for significant differences andTukey’s post hoc to determine where differences occurred (JMP,version 10, SAS Institute Inc.).

Discriminant function analysis was conducted to determinewhether 87Sr:86Sr values could be used to correctly classify indi-vidual fish to the river reach of their capture location (PROCDISCRIM, version 9.3, SAS Institute Inc.). First, 87Sr:86Sr data fromplains killifish were used as a training data set to classify all fishinto one of three isotopically distinct reaches of the River. Thejackknife resampling procedure was used for plains killifish datain a cross-validation technique (one observation is left out) toassess model validity by comparing the predicted with knowncapture location. Second, we classified otolith microchemistryedge readings of Pecos bluntnose shiner to capture locations,which we compared with known capture locations to assess clas-sification success. Third, we classified core and middle 87Sr:86Srvalues to the three isotopically distinct areas to determine move-ment patterns for each Pecos bluntnose shiner. In all cases, weused a probability of 0.5 as the cutoff for classification of location.Unequal within-group covariance was used owing to failure of thetest for equal within-group covariance (�2

2 = 10.14, p = 0.006).

1578 Can. J. Fish. Aquat. Sci. Vol. 72, 2015

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y D

IRE

CT

OR

AT

E O

F C

OL

DW

AT

ER

FIS

HE

RIE

S R

ES

on 1

0/08

/15

For

pers

onal

use

onl

y.

Results

Water chemistry and otolith microchemistry of plainskillifish

Noteworthy differences in 87Sr:86Sr values were observedamong the four water samples throughout the upper Rangelandsand lower Farmlands area of the Pecos River. Three distinct 87Sr:86Sr values were described supporting differences in geologic rocktypes. Values for 87Sr:86Sr in water varied from 0.7082 to 0.7083 inthe upper Rangelands area (Willow and Highway 70, respectively)and 0.7078 (n = 2) in the lower Farmlands area (Dexter and High-way 82; Fig. 2). These four water samples represented only a snap-shot for each location, but were deemed sufficiently different todetermine that a movement study for Pecos bluntnose shiner waswarranted.

Similar to water samples, values for 87Sr:86Sr in otoliths ofplains killifish were reflective of an upper reach (Willow, 6 MileDraw, Crockett Draw, Bosque Draw, Gasline), middle reach (High-way 70), and a lower reach (Dexter, Highway 82) (Fig. 2). In addi-tion, no seasonal shifts in isotopic values of the Pecos River weredetected as indicated by the stability of laser ablation path 87Sr:86Sr values from the primordia to the edge of each plains killifishotolith. Of plains killifish captured above Highway 70 (hereinafterreferred to as the upper reach), 91% (50/55) were classified cor-rectly as originating in that reach, 56% (5/9) were classified cor-rectly to Highway 70, reflecting a transition or mixing zone, andall 33 (100%) plains killifish captured below Highway 70 (herineaf-ter referred to as the lower reach) were classified correctly. Thus,the plains killifish was deemed a suitable surrogate for character-izing Pecos River water 87Sr:86Sr throughout the study area bothspatially (all sites where fish were captured) and temporally(throughout the lives of the fish). As such, additional water sam-ples were not collected to ensure correct classification because ofcost considerations and because plains killifish otolith isotopevalues from the primordia to the edge closely reflected watersamples collected throughout the three distinct isotopic reaches.

Evaluation of Pecos bluntnose shiner classificationsDiscriminant function analysis of otolith 87Sr:86Sr edge values

(i.e., time at capture) indicated that 74% (75/101) of the Pecos blunt-nose shiner were correctly assigned to the upper reach. Twenty-six percent (26/101) of fish were misclassified as captured atHighway 70. Seventy-five percent (3/4) of Pecos bluntnose shinerfrom the lower reach were correctly assigned to the lower reach,while one was misclassified to Highway 70. Only 36% (5/14) of thefish were correctly assigned to Highway 70. Six fish were misclas-sified to the upper reach when they should have been classified toHighway 70. Three fish were misclassified to the lower reach whenthey should have been classified to Highway 70. Classification offish to Highway 70 had greater inherent error as a transition zoneof isotopic values because of the shift in bedrock above and belowthe area.

Reconstructing natal origin and movement of Pecosbluntnose shiner

Otolith primordia or early life 87Sr:86Sr values revealed natalorigins for the majority (82%, 83/101) of Pecos bluntnose shinerbegan in the middle and lower reaches (Table 1). Prior to capture atthe upper-most site (Willow) in the upper reach, 22% (7/32) of Pecosbluntnose shiner began early life in the upper reach, 25% (8/32) ofthese fish began early life in the middle reach, while 53% (17/32)began early life in the lower reach. Thus, 82% of Pecos bluntnoseshiner hatched at, or below, Highway 70, while 18% hatched in theupper reach and remained in the upper reach until capture at Wil-low. Fish captured at Gasline, Highway 70, and Dexter sites con-tained no primordia with 87Sr:86Sr values representative of natalorigins within the upper reach, indicating theses fish began life at, orbelow, Highway 70 (Table 1).

87Sr:86Sr values revealed fewer fish were residents throughouttheir lives within the upper reach. After reconstructing move-ment patterns using otolith 87Sr:86Sr values, the majority of Pecosbluntnose shiner captured at sites within the upper reach hadmoved to the upper reach from either Highway 70 or from the

Fig. 2. Plains killifish (diamonds) were used as a surrogate for identification of shifts in 87Sr:86Sr values from the northern to the southernend of our study area in the Pecos River. Otolith 87Sr:86Sr values from plains killifish revealed a distinct shift at Highway 70. The intermediateand more variable 87Sr:86Sr values observed at Highway 70 indicate a shift in water microchemistry as the bedrock geology changes near thissite. Bars represent mean (±2 standard errors) of 87Sr:86Sr otolith values, and mean 87Sr:86Sr values for plains killifish from upriver anddownriver sites is depicted by the two dashed lines. Circles represent 87Sr:86Sr values of four water samples collected at Willow (0.7083),Highway 70 (0.7082), Dexter (0.7078), and Highway 82 (0.7078).

Chase et al. 1579

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y D

IRE

CT

OR

AT

E O

F C

OL

DW

AT

ER

FIS

HE

RIE

S R

ES

on 1

0/08

/15

For

pers

onal

use

onl

y.

lower reach. Of the total fish captured within the upper reach, 18%(18/101) were life-long residents within the upper reach (Table 2).In contrast, 82% (83/101) of the total fish captured within the upperreach had moved to this reach from either Highway 70 or from thelower reach. 87Sr:86Sr values also revealed large dispersal dis-tances of Pecos bluntnose shiner. For example, the minimumdistance was 56 km from Highway 70 for movement of nonresi-dent fish captured at the Willow site (Table 2).

Age at movementThe majority of Pecos bluntnose shiner captured throughout

the three reaches were 1+ (year class 2011; 78/114) followed by 2+(year class 2010; 35/114). The least abundant age class was 0+ (yearclass 2012) with only two fish captured (Fig. 3). After reconstruct-ing movement patterns using otolith annuli and 87Sr:86Sr valuesfor all age classes, the majority of fish moved upriver within thefirst 90 days of life (Fig. 4). Of the 89 fish captured and classified byage of their first movement either down- or upriver, 74 (83%)moved upriver during summer and fall of their first year (0+early,-mid, -late) and remained in the upper reach until capture (Fig. 5).Of these fish, 41 (46%) had primordia 87Sr:86Sr values that classi-fied to the upper reach, indicating that spawning and hatchoccurred in the upper reach. These fish dispersed downriver post-hatch past Highway 70 to the lower reach and then returned to theupper reach before their first winter. Thus, a little less than half ofPecos bluntnose shiner exhibited large dispersal patterns as evi-dence of three distinct 87Sr:86Sr values reflective of the threereaches. The remaining 48 (54%) had primordia 87Sr:86Sr valuesreflective of the lower reach, whereupon hatch, these fish thenmoved to the upper reach. One exception included a fish withprimordia 87Sr:86Sr values reflective of Highway 70 that moveddownriver from Highway 70, where it was eventually captured atDexter prior to the forming of its second annulus (1+late) for atotal distance of 58 km (Fig. 5).

Swimming performance of Pecos bluntnose shinerSwimming performance tests of Pecos bluntnose shiner sup-

ported our reconstruction of movement from 87Sr:86Sr values inotoliths. Even at an early age of 30 days posthatch, these fishexhibited a strong swimming capacity. Upper critical swimmingspeed (Ucrit) increased with total length, indicating that larger fish

performed better at higher flow rates (Table 3). When consideringsize of fish, higher swimming rate (BL·s−1) was observed in theyoungest fish (30 days posthatch) (Table 3). Total distance swamwas also calculated during swimming trials, revealing fish at30 days posthatch swam a distance of 0.55 km in 83 min and adultfish swam a distance of 1.04 km in 96 min. Water quality waswithin acceptable limits throughout all swimming challenges.Water temperature varied from 19.7 to 20.6 °C, dissolved oxygenvaried from 7.32 to 7.83 mg·L−1, pH varied from 7.51 to 7.74, andconductivity varied from 2.26 to 2.31 mS·cm−1.

DiscussionIsotopic chemistry of the Pecos River across all sampling loca-

tions was successfully characterized both spatially and temporallyusing 87Sr:86Sr values from otoliths of the less mobile plains killi-fish. Strontium values in otoliths of plains killifish revealed thesefish remained within isotopically unique river reaches and thatthere were no detectable seasonal shifts within the study area.Isotopically unique reaches encompassed larger areas than move-ments made by resident plains killifish, thus allowing the useof 87Sr:86Sr in plains killifish otolith as a suitable surrogate for87Sr:86Sr in water. Resident fishes have utility in assessing move-ments of mobile fishes using 87Sr:86Sr values in otoliths especiallyif variability of 87Sr:86Sr values through time can be accounted for(Gillanders 2002). Thus, the mixing zone at Highway 70 repre-sented the confluence of two distinctive bedrock geologies, andwhile this area of transition resulted in less certainty for classify-ing movement, the use of a resident fish improved classificationof movement.

Pecos bluntnose shiner exhibited two patterns of movementthat may explain recruitment into the population. The first morecommon pattern was passive dispersal of propagules downriverbelow Highway 70, followed by movement upriver prior to theirfirst winter (82% of fish captured above Highway 70). The secondless common pattern was passive dispersal with retention of prop-agules in the upriver reach where residents remained at or aboveHighway 70 throughout their lives (18% of the fish captured at orabove Highway 70). Cowley et al. (2009) suggested that bidirec-tional dispersal (downriver displacement and upriver movement)was important in the life history of a similar Great Plains river fish(Rio Grande silvery minnow, Hybognathus amarus). The combinedreturn of young-of-year from downriver and retention as residentssuggests that the current success of the population relies on even-tually securing a large portion of the population through eggretention in the upriver reach. Downriver displacement of prop-agules is advantageous in fishes of Great Plains rivers with vari-able flow regimes and long unobstructed stretches. Movement

Table 1. Percent natal origin of Pecos bluntnose shiner con-structed from otolith 87Sr:86Sr values among three isotopi-cally distinct reaches of the Pecos River (upper, middle, andlower reaches).

Natal origin

Capture site Upper reach Middle reach Lower reach

Upper reachWillow 22% (7/32) 25% (8/32) 53% (17/32)6 Mile 16% (3/19) 47% (9/19) 37% (7/19)Crockett 13% (3/23) 13% (3/23) 74% (17/23)Bosque 22% (5/23) 22% (5/23) 56% (13/23)Gasline 0% (0/4) 25% (1/4) 75% (3/4)

Middle reachHwy 70 0% (0/14) 21% (3/14) 79% (11/14)

Lower reachDexter 0% (0/4) 25% (1/4) 75% (3/4)

Note: Numerators in parentheses represent the number of fishwith primordia 87Sr:86Sr values classified to one of the three reachesindicating natal origins within the reach. Denominators are total fishcaptured at one of the five sites in the upper reach (five sites, n = 101),middle reach at Highway 70 (n = 14), or lower reach at Dexter (n = 4).Capture sites are listed from upriver to downriver (i.e., Willow is themost northern site in the upper reach, and Dexter is the most south-ern site in the lower reach) and represent where fish were captured inrelation to the reach of their natal origin.

Table 2. Resident and upriver movement of Pecos bluntnose shinerbefore capture within the upriver reach (above Highway 70).

Capturesite

Residentswithinupper reach

Movement upriverfrom Highway 70 orfrom lower reach

Distance moved upriverfrom Highway 70(river kilometres)

Willow 22% (7/32) 78% (25/32) 566 Mile 16% (3/19) 84% (16/19) 45Crockett 13% (3/23) 87% (20/23) 37Bosque 22% (5/23) 78% (18/23) 27Gasline 0% (0/4) 100% (4/4) 11

Total 18% (18/101) 82% (83/101) —

Note: Capture site represents capture location of fish. Percentage of residentsabove Highway 70 within the upper reach is compared with those that movedupriver from either Highway 70 or from the lower reach until capture. Numer-ators in parentheses are the number of fish expressing 87Sr:86Sr values reflectingeither resident of the upper reach or movement upriver to the upper reach,while denominators are the total fish captured at each site. Distance moved(river kilometres) of Pecos bluntnose shiner is depicted to reflect minimumdistance of movement from Highway 70 to the capture site.

1580 Can. J. Fish. Aquat. Sci. Vol. 72, 2015

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y D

IRE

CT

OR

AT

E O

F C

OL

DW

AT

ER

FIS

HE

RIE

S R

ES

on 1

0/08

/15

For

pers

onal

use

onl

y.

upriver for spawning by adult fish ensures adequate distance foregg drift and development downriver (Cross et al. 1985; Durhamand Wilde 2008). However, discharge, habitat complexity, andfragmentation ultimately affect egg retention. Using gellan beadsas egg surrogates, Dudley and Platania (2007) demonstrated trans-port distances of propagules from pelagic broadcast spawningfishes might occur three times further during sustained reservoirreleases. Additionally, Worthington et al. (2014) demonstrateddownstream drift of egg surrogates was driven by increased veloc-ity but that as habitat complexity increased, a larger proportion of

eggs were retained. Within the Pecos River, if propagules are notdeposited into slack-water nursery habitat, then young fish areeventually at risk of downriver displacement and removal fromthe population (Brooks et al. 1994; Platania and Altenbach 1998).Hoagstrom et al. (2008a) related high densities of young Pecosbluntnose shiner at the inflow of Brantley Reservoir to a longrelease of water from Sumner Dam.

Successful fish movement upriver and maintaining position inflow rely on fish size and behavior (Ward et al. 2003) as well asflow velocity and channel morphology (Leavy and Bonner 2009).

Fig. 3. Age distribution (number of fish) of Pecos bluntnose shiner for all age classes and capture locations shows that the majority of age 1+and 2+ are found throughout the upper reaches of the Pecos River. Sample size is in parentheses.

Fig. 4. 87Sr:86Sr values from the primordia to the edge of an individual Pecos bluntnose shiner that dispersed from the lower reach to theupper reach of the Pecos River during its lifetime.

Chase et al. 1581

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y D

IRE

CT

OR

AT

E O

F C

OL

DW

AT

ER

FIS

HE

RIE

S R

ES

on 1

0/08

/15

For

pers

onal

use

onl

y.

Pecos bluntnose shiner prefer shallower depths coupled with rel-atively swift water velocity typical of wide shifting sand-bed rivers(Hoagstrom et al. 2008a). Hoagstrom et al. (2008b) associatedlength classes of Pecos bluntnose shiner with water velocity andfound a positive relationship between increasing water velocityand fish size. Upper critical swimming speeds (Ucrit, 43 cm·s−1) forPecos bluntnose shiner as young as 30 days posthatch revealed theupper threshold of aerobic swimming capacity is high for thisspecies at an early age. Caldwell et al. (2010) reported 34.3–44.1 cm·s−1 for upper critical swimming speed in Rio Grandesilvery minnow at 116 days posthatch. In comparison, Pecosbluntnose shiner exhibited higher swimming capacity (Ucrit,52.5 cm·s−1) at 90 days posthatch. Bestgen et al. (2010) reportedupper critical swimming speeds of 51.5 cm·s−1 (53–75 mm TL) forRio Grande silvery minnow, noting that swimming abilityincreased with fish size. Although Pecos bluntnose shinerdemonstrated strong swimming ability and upriver movement,downriver dispersal of propagules has potentially tripled to dis-tances up to 142 km from predam–prechannelization of the river(Dudley and Platania 2007).

Persistence of Pecos bluntnose shiner relies on a variety of abi-otic factors that include timing of pulse flows that cue spawningevents, habitat quality, and perennial flow to maintain river con-nectivity for bidirectional dispersal to complete their life cycle.Using otolith microchemistry, we recorded that movement in thepopulation coincided with years of perennial flow throughoutsummer and fall (fish hatched 2010 and 2011) before the onset oftheir first winter. During the study (2011–2012), the Southwestexperienced one of the most severe droughts on record (http://www.droughtmonitor.unl.ed-u/archive.html, accessed 10 November2012). Our capture and otolith analysis of only two 2012 age class0+ fish was supported by Davenport (2012), who reported less than1% of age class 0+ fish in 2012 due to river intermittency fromdrought. From 17 July to 20 August 2012, 55 km of the upperRangelands area of the Pecos River dried and presumably not only

halted fish movement but limited recruitment into the popula-tion (S. Davenport, United States Fish and Wildlife Service, Albu-querque, New Mexico, personal communication, 2012). Ourotolith analysis revealed that when river conditions supportedfish spawning and movement, the 2008 age class 0+ represented16%–51% of Pecos bluntnose shiner collected throughout thereaches of the river.

Otolith 87Sr:86Sr values revealed new information on life his-tory movement patterns of a small-bodied but highly mobile fish.As a relatively short-lived species, Pecos bluntnose shiner mustmove upriver early in life such that when the opportunity tospawn occurs, propagules have sufficient distance to drift whiledeveloping. Swimming performance tests confirmed that youngPecos bluntnose shiner were capable of moving upriver prior totheir first winter where they gain reproductive advantage of thedispersal distance for propagule development. These movementpatterns represent a trade-off between years with stream dryingin upper sections and the minimum distance needed for adequatedevelopment (i.e., bet-hedging strategy). Habitat restoration inthe southern Farmlands area of the Pecos River would likely ben-efit pelagic spawning fish by returning this area from a highlymodified channel to braided, more complex areas with shiftingsand-bed. This would increase backwater areas for nursery habitatand thereby increase chances of success for the return of the Pecosbluntnose shiner to the upper reach and completion of its lifehistory.

AcknowledgementsFunding was provided by US Fish and Wildlife Service – Science

Support Partnership (M. Mata-Gonzalez and S. Oetker). Field assis-tance was provided by the US Fish and Wildlife Service – Fish andWildlife Natural Resource Conservation Office (S. Davenport andS. Blocker). Pecos bluntnose shiner and access to the swim tunnelwas provided by US Fish and Wildlife Service – Southwestern Na-tive Aquatic Resources and Recovery Center, Dexter, New Mexico(M. Ulibarri, W. Knight, C. Sykes). J. Glessner, G. Barford, andJ. Commisso of the University of California-Davis provided assis-tance with isotopic analysis. Invaluable insight of the ecology ofpelagic spawning fishes was provided by C. Hoagstrom, S. Platania,S. Brewer, and D. Cowley. S. Brewer provided a substantivereview of the manuscript. Any use of trade, firm, or productnames is for descriptive purposes only and does not imply en-dorsement by the US Government. This study was performed un-der the auspices of New Mexico State University’s InstitutionalAnimal Care and Use Committee (No. 2012-019). The manuscriptwas improved by two anonymous reviewers.

Fig. 5. Number and timing of movements by Pecos bluntnose shiner dispersing downriver and upriver reveal that upstream movementoccurred in the first year. 0+ have not formed an annulus, 1+ have one annulus, and 2+ have two annuli. Sample size is in parentheses.

Table 3. Mean total length (TL, mm), critical swimming speed (Ucrit,cm·s–1), swimming rate (body lengths (BL)·s–1), and mean total distanceswam (km) of four age classes of Pecos bluntnose shiner.

Ageclass TL (mm) Ucrit (cm·s–1) BL·s–1

Totaldistance (km)

30-day 21.3 (0.62) 43.8 (4.46) 20.6 (2.02) 0.55 (0.094)60-day 33.9 (0.92) 49.2 (1.94) 14.5 (0.52) 0.62 (0.031)90-day 46.3 (0.82) 52.5 (2.48) 11.3 (0.62) 0.68 (0.042)Adult 69.1 (2.28) 70.3 (3.26) 10.2 (0.54) 1.04 (0.074)

Note: Values in parentheses are two standard errors. Sample size of 30 fishwas used for 30-day, 60-day, and adult age classes, while 15 fish were tested for90-day age class.

1582 Can. J. Fish. Aquat. Sci. Vol. 72, 2015

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y D

IRE

CT

OR

AT

E O

F C

OL

DW

AT

ER

FIS

HE

RIE

S R

ES

on 1

0/08

/15

For

pers

onal

use

onl

y.

ReferencesBarnett-Johnson, R., Pearson, T.E., Ramos, F.C., Grimes, C.B., and

MacFarlane, R.B. 2008. Tracking natal origins of salmon using isotopes, oto-liths, and landscape geology. Limnol. Oceanogr. 53: 1633–1642. doi:10.4319/lo.2008.53.4.1633.

Beamish, F.W.H. 1978. Swimming capacity. In Fish physiology volume VII: Loco-motion. Edited by W.S. Hoar and D.J. Randall. Academic Press, New York.pp. 101–187.

Bestgen, K.R., and Platania, S.P. 1990. Extirpation of Notropis simus simus (Cope)and Notropis orca (Woolman) (Pisces: Cyprinidae) from the Rio Grande in NewMexico, with notes on their life history. Museum of Southwestern Biology,6: 1–8.

Bestgen, K.R., Platania, S.P., Brooks, J.E., and Propst, D.L. 1989. Dispersal and lifehistory traits of Notropis girardi (Cypriniformes: Cyprinidae) introduced intothe Pecos River, New Mexico. Am. Midl. Nat. 122: 228–235. doi:10.2307/2425907.

Bestgen, K.R., Mefford, B., Bundy, J.M., Walford, C.D., and Compton, R.I. 2010.Swimming performance and fishway model passage success of Rio Grandesilvery minnow. Trans. Am. Fish. Soc. 139(2): 433–448. doi:10.1577/T09-085.1.

Brooks, J.E., Brown, M.R., and Allan, N.L. 1994. Pecos River fishery investigations:fish community structure, habitat use and water quality as a response toreservoir operations. 1992 Annual Report. US Bureau of Reclamation. Albu-querque, New Mexico.

Caldwell, C.A., Barrows, F.T., Ulibarri, M., and Gould, W.R. 2010. Diet optimiza-tion of juvenile Rio Grande silvery minnow. North Am. J. Aquat. 72(1): 57–64.doi:10.1577/A09-011.1.

Cowley, D.E., Alleman, J.C., Sallenave, R., McShane, R.R., and Shirey, P.D. 2009.Effects of salinity on specific gravity and viability of eggs of a North Americanminnow (Cyprinidae). Sci. Mar. 73: 47–58. doi:10.3989/scimar.2009.73s1047.

Cross, F.B., Moss, R.E., and Collins, J.T. 1985. Assessment of dewatering impactson stream fishes in the Arkansas and Cimarron rivers. Kansas Fish and GameCommission Nongame Contract, 46: 1–161.

Davenport, S.R. 2010. Status and trends of Pecos bluntnose shiner Notropis simuspecosensis Pecos River, New Mexico. Report 1-34. US Bureau of Reclamation,Albuquerque, New Mexico.

Davenport, S.R. 2012. Status and trends of Pecos bluntnose shiner Notropis simuspecosensis Pecos River, New Mexico. US Bureau of Reclamation, AlbuquerqueNew Mexico.

Dodds, W.K., Gido, K., Whiles, M.R., Fritz, K.M., and Matthews, W.J. 2004. Life onthe edge: the ecology of Great Plains prairie streams. BioScience, 54: 205–216.doi:10.1641/0006-3568(2004)054[0205:LOTETE]2.0.CO;2.

Dudley, R.K., and Platania, S.P. 2007. Flow regulation and fragmentation imperilpelagic-spawning riverine fishes. Ecol. Appl. 17: 2074–2086. doi:10.1890/06-1252.1.

Durham, B.W., and Wilde, G.R. 2005. Relationship between hatch date andfirst-summer growth of five species of prairie-stream cyprinids. Environ. Biol.Fishes, 72: 45–54. doi:10.1007/s10641-004-4186-5.

Durham, B.W., and Wilde, G.R. 2008. Asynchronous and synchronous spawningby smalleye shiner Notropis buccula from the Brazos River, Texas. Ecol. Freshw.Fish, 17: 528–541. doi:10.1111/j.1600-0633.2008.00303.x.

Durham, B.W., and Wilde, G.R. 2009. Effects of streamflow and intermittency onthe reproductive success of two broadcast-spawning cyprinid fishes. Copeia,1: 21–28. doi:10.1643/CE-07-166.

Gillanders, B.M. 2002. Temporal and spatial variability in elemental composi-tion of otoliths: implications for determining stock identity and connectivityof populations. Can. J. Fish. Aquat. Sci. 59: 669–679. doi:10.1139/f02-040.

Hatch, M.D., Baltosser, W.H., and Schmitt, C.G. 1985. Life history and ecology ofthe Bluntnose Shiner (Notropis simus pecosensis) in the Pecos River of NewMexico. Southwest. Nat. 30(4): 555–562. doi:10.2307/3671049.

Hegg, J., Kennedy, B., Chittaro, P., and Zabel, R. 2013. Spatial structuring of anevolving life-history strategy under altered environmental conditions. Oeco-logia, 172: 1017–1029. doi:10.1007/s00442-012-2564-9. PMID:23423520.

Hoagstrom, C.W. 2003a. Pecos Bluntnose Shiner suitability, Pecos River, NewMexico 1992–1999. US Bureau of Reclamation. Albuquerque, New Mexico.

Hoagstrom, C.W. 2003b. Pecos bluntnose shiner population dynamics and sea-sonal flow regime, Pecos River, New Mexico. US Bureau of Reclamation.Albuquerque, New Mexico.

Hoagstrom, C.W. 2009. Causes and impacts of salinization in the lower PecosRiver [online]. Gt. Plains Res.: A Journal of Natural and Social Sciences, 19:27–44. Available from http://digitalcommons.unl.edu/greatplainsresearch/994.

Hoagstrom, C.W., and Turner, T.F. 2015. Recruitment ecology of pelagic-broadcast spawning minnows: paradigms from the ocean advance scienceand conservation of an imperiled freshwater fauna. Fish Fish. 16(2): 282–299.doi:10.1111/faf.12054.

Hoagstrom, C.W., Brooks, J.E., and Davenport, S.R. 2008a. Spatiotemporal pop-ulation trends of Notropis simus pecosensis in relation to habitat conditions andthe annual flow regime of the Pecos River, 1992–2005. Copeia, 1: 5–15. doi:10.1643/CE-07-002.

Hoagstrom, C.W., Brooks, J.E., and Davenport, S.R. 2008b. Recent habitat asso-ciation and the historical decline of Notropis simus pecosensis. River Res. Appl.24: 789–803. doi:10.1002/rra.1082.

Hoagstrom, C.W., Brooks, J.E., and Davenport, S.R. 2011. A large-scale conserva-tion perspective considering endemic fishes of the North American plains.Biol. Conserv. 144: 21–34. doi:10.1016/j.biocon.2010.07.015.

Hobbs, J.A., Lewis, L.S., Ikemiyagi, N., Sommer, T., and Baxter, R.D. 2010. The useof otolith strontium isotopes (87Sr/86Sr) to identify nursery habitat for athreatened estuarine fish. Environ. Biol. Fishes, 89: 557–569. doi:10.1007/s10641-010-9672-3.

Leavy, T.R., and Bonner, T.H. 2009. Relationships among swimming ability, cur-rent velocity association, and morphology of freshwater lotic fishes. NorthAm. J. Fish. Manage. 29: 72–83. doi:10.1577/M07-040.1.

Miller, S.J., and Storck, T. 1984. Temporal spawning distribution of largemouthbass and young-of-year growth, determined from daily otolith rings. Trans.Am. Fish. Soc. 113(5): 571–578. doi:10.1577/1548-8659(1984)113<571:TSDOLB>2.0.CO;2.

Minckley, C.O., and Klaassen, H.E. 1969. Life history of the plains killifish, Fun-dulus kansae (Garman), in the Smoky Hill River, Kansas. Trans. Am. Fish. Soc.98: 460–465. doi:10.1577/1548-8659(1969)98[460:LHOTPK]2.0.CO;2.

NMBGR. 2003. Geologic Map of New Mexico. 1:500,000. New Mexico Bureau ofGeology and Mineral Resources.

NMDGF. 2012. Threatened and endangered species of New Mexico 2012 BiennialReview. New Mexico Department of Game and Fish, Santa Fe, New Mexico.

Platania, S.P. 1995. Distribution, relative abundance, and conservation status ofPecos bluntnose shiner (Notropis simus pecosensis). New Mexico Department ofGame and Fish, Santa Fe, New Mexico.

Platania, S.P., and Altenbach, C.S. 1998. Reproductive strategies and egg type ofseven Rio Grande Basin cyprinids. Copeia, 3: 559–569. doi:10.2307/1447786.

Tashjian, P.L. 1993. Channel response, reservoir operations and Pecos BluntnoseShiner habitat, Pecos River, NM. 1992 Annual Progress Report. US Bureau ofReclamation, Albuquerque, New Mexico.

Thorrold, S.R., Jones, C.M., Swart, P.K., and Targett, T.E. 1998. Accurate classifi-cation of juvenile weakfish Cynoscion regalis to estuarine nursery areas basedon chemical signatures in otoliths. Mar. Ecol. Prog. Ser. 173: 253–265. doi:10.3354/meps173253.

USFWS. 1987. Endangered and threatened wildlife and plants; determination ofthreatened status for the Pecos bluntnose shiner designation of its criticalhabitat. Fed. Regist. 52(B4): 5295–5303. US Fish and Wildlife Service.

Walker, J.D., and Geissman, J.W. 2009. Geologic time scale: Geological Society ofAmerica. Geol. Soc. Am. doi:10.1130/2009.CTS004R2C.

Ward, D.L., Schultz, A.A., and Matson, P.G. 2003. Differences in swimmingability and behavior in response to high water velocities among nativeand nonnative fishes. Environ. Biol. Fishes, 68: 87–92. doi:10.1023/A:1026031128486.

Worthington, T.A., Brewer, S.K., Farless, N., Grabowski, T.B., and Gregory, M.S.2014. Interacting effects of discharge and channel morphology on transportof semibuoyant fish eggs in large, altered river systems. PLoS One, 9(5):e96599. doi:10.1371/journal.pone0096599. PMID:24802361.

Chase et al. 1583

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y D

IRE

CT

OR

AT

E O

F C

OL

DW

AT

ER

FIS

HE

RIE

S R

ES

on 1

0/08

/15

For

pers

onal

use

onl

y.