3Centre for Conservation Ecology and Environmental Science, Bournemouth University, BH12 5BB Poole, United Kingdom
4Environmental and Life Sciences Graduate Program, Trent University, ON K9J 7B8 Peterborough, Canada
Lorenzo Vilizzi1,*, Gordon H. Copp2,3,4
Bias, precision and validation of ageing 0+ European barbel Barbus barbus (L.) from
their otoliths
1. IntroductionAge and growth are key aspects of fish biology with which to inform fisheries managers on the general well-being and status of a stock, and this is particularly crucial in studies of growth in early life, which is a major determinant of over-winter survival [1,2]. Although classed in the IUCN Red List as being of ‘Least Concern’, the European barbel Barbus barbus (L.) is considered a ‘flag’ species for river conservation due to its sensitivity to pollution and its popularity, and thus socio-economic importance, as a sport fish [3]. Increasingly threatened in parts of its native range in Central Europe [4], B. barbus populations are endangered locally by water pollution and river regulation, including in the U.K., where the
species is native to eastern rivers between Yorkshire and the Thames [5]. Apparent declines in population densities, such as in the River Lee, Hertfordshire [6,7], have emphasised the need for accuracy (hence, validation) and precision in the ageing of 0+ (young-of-year) B. barbus so as to improve understanding of growth rates during early life for appropriate management and conservation of B. barbus stocks [3,8].
As part of a broader investigation of the environmental biology of B. barbus early life stages in the River Lee [7,9-15], the aim of the present study was to evaluate bias and precision of otolith (daily) micro-increment counts (sensu Campana and Neilson [16]) in 0+ B. barbus and to validate the age determinations. The specific objectives were to: (i) determine which of
Abstract: The European barbel Barbus barbus L. is considered a ‘flag’ species for river conservation and sport fishing, but it is increasingly threatened in its native range of distribution. To provide accurate age estimates during early life for appropriate management and conservation measures, the bias and precision of otolith (daily) micro-increment counts were evaluated and age determinations validated on laboratory-reared embryos and larvae. Out of the three pairs of otoliths, the lapillus and sagitta provided reliable age estimates for free embryos and larvae up to 17 days of (known) age post-fertilisation, with first micro-increment formation occurring five days post-fertilisation. On the other hand, micro-increments on asterisci formed only 16–17 days post-fertilisation. There was agreement in micro-increment counts based on lapilli and sagittae, but not between interpreters, indicating that despite consistency between the two pairs of otoliths extensive training and experience are required for reliable age interpretation. The ability to estimate the ages of 0+ B. barbus from their otoliths will contribute to a better understanding of growth rates from both hatchery-stocked and native/introduced cohorts.
the three pairs of otoliths can be used more reliably for age estimation based on micro-increment counts; (ii) assess bias and precision of micro-increment counts both between and within structures (i.e. otolith pairs) as well as between interpreters; and (iii) establish a relationship between known (post-fertilisation) age and otolith-based micro-increment counts for 0+ B. barbus in their first days of life (cf. validation).
2. Experimental Procedures2.1 Sample collection and preparationFertilised B. barbus eggs were obtained from the artificial spawning of wild-captured male and female individuals (female SL: n=5, mean =392.8±18.7 SE mm, min–max =345–460 mm; female W: n=5, mean =852±105.2 SE g, min–max =600–1230; male SL: n=7, mean =359.6±9.80 SE mm, min–max =324–395 mm; male W: mean =687.1±46.6 SE g, min–max =500–800 g) as described in Vilizzi and Copp [15] and placed in a plastic, fine-mesh basket suspended within a 60 L glass tank. Water in the tank was drawn from domestic water supply and was recirculated through an experimental flume system for at least two weeks prior to the study to ensure thorough de-chlorination and ‘maturation’ (see Vilizzi and Copp [15]). Light in the room, which had a large bay window, was determined by natural day length and sun intensity, with water temperature maintained at 21±1°C. Following emergence on Day 6 (cf. Peňáz [17]; Vilizzi and Copp [15]), the larvae were fed ad libitum brine shrimp nauplii (Artemia spp.).
Prior to dissection, each fish was measured for egg diameter (embryos) and standard length (SL; free embryos and larvae) to the nearest 0.01 mm using dial callipers under a binocular dissecting microscope at 45× (embryos) and 30× (free embryos and larvae)—terminology and timing of developmental intervals after Peňáz [17] and Vilizzi and Copp [15], respectively. From Day 1 (post-fertilisation: 12:00 p.m.) until Day 15, and then on Day 17 and on Day 20 (one fish only), eight to eleven randomly-sampled fish were killed each day with an overdose of anaesthetic and severance of the head. Notably, observations in the present study ended on Day 20 as this marked the transition from the larval to the juvenile period at current temperature conditions [17]. The entire set of otoliths was excised, with the lapilli (utricular otoliths), sagittae (saccular otoliths) and asterisci (lagenar otoliths) identified [18]. Polarised light was used by a home-made adaptation of a linear polarising photo filter on both microscopes to increase the contrast between calcified structures and organic
tissues. Following extraction, each otolith was first cleaned of any surrounding tissue (using needles under non-polarised light) and placed on a labelled glass slide, which was then heated on a hot plate (100–130°C). The otolith was finally mounted horizontally using a drop of Crystalbond (Aremco, Valley Cottage, NY, USA) under a 400× compound microscope. Micro-increments were identified under transmitted light as bipartite structures, each composed of an incremental (translucent) and a discontinuous (opaque) zone [16], and no polishing was required for reliable micro-increment counts due to the early age of the fish under study. Micro-increment counts were made independently by two interpreters (A and B) without any knowledge of either fish age or size after the otolith extraction process was completed.
2.2 Data analysisBias and precision of micro-increment counts were analysed by statistical and graphical methods for three pairs of combinations: (i) between-structure and within-interpreter (i.e. lapillus vs sagitta from Interpreter A; lapillus vs sagitta from Interpreter B); (ii) within-structure and between-interpreter (i.e. lapillus from Interpreter A vs B; sagitta from Interpreter A vs B); and (iii) between-structure and between-interpreter (i.e. lapillus from Interpreter A vs sagitta from Interpreter B; sagitta from Interpreter A vs lapillus from Interpreter B). For each combination, an age-bias plot [19] was generated and a test of symmetry [20] computed from the corresponding age-agreement table (not shown). In this table, the main diagonal represents the frequency of fish for which the same number of micro-increments is obtained based on the two structures and/or interpreters, whereas each cell off the main diagonal represents a difference in the number of micro-increments between the two structures and/or interpreters. If there are no systematic differences between structures and/or interpreters, then it is expected that the disagreements in the number of micro-increments would fall randomly on either side of the diagonal, producing an approximately symmetric age-agreement table. Precision, defined as the reproducibility of micro-increment counts between structures and/or interpreters [21], was computed as percentage agreement, average percent error (APE), and by the coefficient of variation (CV). All computations were in the R language and software environment for statistical computing and graphics v2.13.0 64-bit [22] using libraries FSA and NCStats and following guidelines in Ogle [23].
To test for overall differences in micro-increment counts from the first and second lapillus and sagitta in a pair examined by the two interpreters, permutational univariate analysis of variance (PERANOVA) was
Bias, precision and validation of ageing 0+ European barbel Barbus barbus (L.) from their otoliths
employed on a three-factor design consisting of factors Pair (first, second), Otolith (lapillus, sagitta) and Interpreter (A, B), all fixed and crossed. Data (i.e. micro-increment counts) were first normalised and a Euclidean distance dissimilarity matrix was obtained, with statistically significant effects at α=0.05. Computations were in PERMANOVA+ v1.0.1 for PRIMER v6.1.11, with 9999 permutations of the raw data [24].
Finally, a simple linear regression model was fitted to the relationship between known (post-fertilisation) age (days) and the estimated age based on micro-increment counts separately for lapilli and sagittae. Fitting and related statistics were in Excel® 2010 for Windows using the Analysis ToolPak add-in.
3. ResultsOtoliths were examined from a total of 159 fish, comprising 49 embryos (age 1–5 days post-fertilisation; diameter =2.75±0.08 SE mm, min–max =
2.66–2.96 mm), 61 free embryos (6–11 days; SL=10.68±0.85 SE mm, min–max =9.10–12.24 mm) and 49 larvae (12–20 days; SL=12.68±0.50 SE mm, min–max =12.00–14.30 mm). Lapilli were first detected in embryos four days post-fertilisation and were present in all dissected free embryos after six days; sagittae also were detected four days post-fertilisation and were present in all dissected free embryos after seven days. On the contrary, asterisci were found to have formed only after 17 days post-fertilisation (with none found up to 15 days) and were present in all dissected larvae after 20 days (Table 1). Owing to the late formation of micro-increments on the asterisci, the lapilli and sagittae from fish 1–20 days post-fertilisation were examined for micro-increment counts.
In total, 223 lapilli and sagittae (105 pairs and 13 odd, in both cases) were extracted. The entire complement was examined for micro-increment counts by Interpreter A, whereas a subset consisting of 175 lapilli (84 pairs and 7 odd) and 160 sagittae (75 pairs and 10 odd) was examined by interpreter B. The proportion of lapilli and
Otolith
Size (mm)1 Lapillus Sagitta Asteriscus
Developmental interval
Age (days) n Mean SE n (no, np) % Fish n (no, np) % Fish n (no, np) % Fish
Table 1. Laboratory-reared 0+ European barbel Barbus barbus examined for otolith (daily) micro-increment counts. The age from fertilisation is given for the three developmental intervals under study (after Peňáz [17] and Vilizzi and Copp [15]) along with size (± SE: standard error), number of otoliths (n) recovered (with odd otoliths no and paired otoliths np in parentheses, the sum of which equals the number of fish with otoliths recovered from) along with percentage of fish with recovered otoliths.
1Egg diameter for embryos and standard length for free embryos and larvae. 2Proportion <100% due to unrecovered otoliths.
sagittae discarded as uninterpretable by Interpreter A was 7.2% and 7.7%, respectively; whereas for Interpreter B this proportion was 15.6% and 15.2%, respectively. Micro-increments were first detected in both lapilli and sagittae five days post-fertilisation, hence with no (visible) micro-increments at the age of four days. Pair-wise bias comparisons indicated overall similarities in micro-increment counts within interpreters (Figure 1a,b), between interpreters relative to lapilli (Figure 1c), and between sagittae examined by Interpreter A and lapilli examined by Interpreter B (Figure 1f).
On the other hand, there was a tendency for interpreter A to underestimate micro-increment counts on sagittae beyond six increments relative to Interpreter B (Figure 1d), and on lapilli beyond five increments relative to sagittae examined by Interpreter B (Figure 1e), even though bias was statistically significant only in the former case (Table 2). Agreement was higher within Interpreter B (i.e. lapilli vs sagittae) and this was reflected by the lower corresponding APE and CV values, which in all other cases were close to 10 or higher (Table 2). Overall, there were no statistically significant differences in mean micro-increments counts between lapilli and sagittae or between the first and second otolith in a pair. However, significant differences were found between interpreters, due to Interpreter A underestimating overall micro-increment counts relative to Interpreter B (7.6±0.3 SE vs 8.5±0.3 SE, respectively; n=156; Table 3).
Given the above discrepancies, validation of ageing was based on micro-increment counts from lapilli and sagittae (averaged over a pair, whenever available) examined by Interpreter B in virtue of his more extensive experience in otolith micro-increment ageing, with the only exception for micro-increment counts on fish 1–5
days of age that were not examined by Interpreter B and were therefore added to the dataset based on counts from Interpreter A. Notably, inclusion of these micro-increment counts from Interpreter A was justified given the aforementioned overall lack of between-interpreter discrepancy below five to six micro-increments.
There was a highly significant linear relationship between known age and number of micro-increment counts from both lapilli (n=92; a=4.703±0.201 SE, b=0.814±0.023 SE; r2=0.932; P<0.001) and sagittae (n=87; a=4.452±0.202 SE, b=0.787±0.022 SE; r2=0.940; P<0.001), with the corresponding equations indicating that the first (identifiable) micro-increment in both the lapillus and the sagitta appears four to five days post-fertilisation, hence with no (identifiable) micro-increments up to three days post-fertilisation (Figure 2).
4. DiscussionThe present study is the first to validate the ageing of 0+ B. barbus by means of otolith micro-increment counts, and to assess related bias and precision by evaluation of the entire set of otoliths to determine their comparative value for age interpretation [25]. The present findings indicate that, based on micro-increment counts, both the lapillus and the sagitta can be used reliably to estimate the ages of 0+ B. barbus at least until the age of 20 days. This contrasts the asterisci, which do not appear until well into the larval period (Table 1). Appearance of lapilli and sagittae four days post-fertilisation is in accordance with the reported ‘otolith’ formation during the embryo phase of development at comparable ages [17,26], even though no indication is provided in the cited studies on the specific type of otolith pair examined.
Bias Precision
Combination n df χ2 P Agreement APE CV
Between-structure and within-interpreter
Lapillus A vs Sagitta A 103 23 33.57 0.072 39.80 7.32 10.36
Lapillus B vs Sagitta B 70 21 25.00 0.247 48.57 3.80 5.38
Within-structure and between-interpreter
Lapillus A vs Lapillus B 78 29 31.00 0.365 34.61 8.46 11.96
Sagitta A vs Sagitta B 76 25 39.13 0.036 31.58 6.95 9.83
Between-structure and between-interpreter
Lapillus A vs Sagitta B 74 30 42.33 0.067 31.08 8.39 11.87
Sagitta A vs Lapillus B 77 24 35.81 0.057 29.87 6.87 9.72
Table 2. Bias (test of symmetry) and precision of otolith micro-increment counts on laboratory-reared 0+ Barbus barbus for three pairs of combinations of otolith (lapillus, sagitta) and Interpreter (A, B). APE = Average Percentage Error; CV = Coefficient of Variation. Significant P values (α=0.05) in bold. See also Figure 1.
Bias, precision and validation of ageing 0+ European barbel Barbus barbus (L.) from their otoliths
Figure 1. Age-bias plots for otolith micro-increments counts on laboratory-reared 0+ European barbel Barbus barbus based on three pairs of combinations of otolith (lapillus, sagitta) and Interpreter (A, B). Crosses represent the mean micro-increment count for each x-axis value.
658
Author's co
py
L. Vilizzi, G.H. Copp
The unsuitability of the asteriscus for micro-increment counts in 0+ B. barbus is consistent with findings for other Eurasian cyprinids such as the Andalusian barbel Luciobarbus sclateri [18] and common carp Cyprinus carpio carpio [27], even though for these species only the lapilli (hence, not the sagittae) have so far been identified as suitable for the ageing of their 0+ individuals [27,28], and this appears to be true also of other North American cyprinids [29]. On the other hand, in a comparative ageing study on the otoliths of three cyprinids from Lake Malawi [30], up to 30 increments on the sagittae could be counted reliably and the it was noted that on the asterisci the number of increments was much lower compared to the lapilli (the most reliable otolith pair for ageing those species). These observations corroborate the findings of the present study, with micro-increments in 0+ B. barbus asterisci
Source df MS F#/t P
Pair 1 0.023 0.02 0.878
Otolith 1 0.957 0.95 0.330
Interpreter 1 5.117 5.10 0.025
Pair × Otolith 1 0.001 0.01 0.974
Pair × Interpreter 1 0.023 0.02 0.877
Otolith × Interpreter 1 0.015 0.01 0.900
Pair × Otolith × Interpreter 1 0.008 0.01 0.922
Residual 304 1.003
Table 3. Differences in micro-increment counts from the first and second (Pair) of lapillus and sagitta (Otolith) from laboratory-reared 0+ Barbus barbus examined by two interpreters (Interpreter). Statistically significant effects for PERANOVA are at α=0.05 (in bold). F# = permutational F-value.
Figure 2. Linear regression model for the relationship between known (post-fertilisation) age (days) and the estimated age of laboratory-reared 0+ Barbus barbus based on micro-increment counts separately for lapillus and sagitta.
Bias, precision and validation of ageing 0+ European barbel Barbus barbus (L.) from their otoliths
appearing several days after lapilli and sagittae, and with identifiable micro-increments on the latter otolith pair falling within the time span of the present study. Therefore, it may be speculated that because of the morphological features of sagittae in cyprinids (e.g. [18,30]), similar to L. sclateri, these otoliths may not be suitable for the ageing of 0+ B. Barbus after their first few weeks of life.
The lack of significant differences in micro-increment counts both between/within structures and between/within interpreters, but with the notable exception of the sagittae examined by the two interpreters, emphasizes the reliability of age estimates from lapilli and sagittae especially when examined by a single interpreter. On the other hand, the significant difference found between the two interpreters, which became more evident upon comparison of their mean micro-increment counts, indicates that extensive training and experience are required for reliable and consistent age interpretation (e.g. [31]).
The formation of otoliths in B. barbus approximately four days following fertilisation as reported in early development studies (see above) and the detection of micro-increments at the age of five days suggest that estimated ages based on micro-increment counts from lapilli and sagittae may only be biased by one increment, should the ‘resolution hypothesis’ (cf. [32]) hold also for 0+ B. barbus ageing. According to this hypothesis, daily increments narrower than the functional resolution limit for light microscopy may form in some 0+ fish species, so that age estimates from micro-increment counts may become conservative and therefore require
higher-resolution scanning electron microscopy (SEM) [28]. In the present study, light microscopy was the only feasible option, so further studies are required to prove or disprove the above hypothesis. Additionally, validation beyond the range of ages examined in the present study (i.e. 0–17) on wild-captured juveniles will be required to assess the extent of the age interval that can be validated under laboratory conditions, which are likely to differ from those in the natural environment.
Stocking of hatchery-reared B. barbus into lentic fisheries in England for the enhancement of catch-and-release angling has been common practice during the last twenty years [33], with recent laboratory-based research focussing on the growth rates of known-age 0+ individuals under different conditions of inter- and intra-specific competition [8]. The ability to estimate reliably the ages of 0+ B. barbus from their otoliths could, therefore, prove to be of additional value for determining the growth rates of cohorts from both hatchery-stocked and native/introduced 0+ B. barbus so as to assess recruitment success and survival beyond the first year of life.
AcknowledgementsThis work was funded through a Marie Curie post-doctoral fellowship from the European Commission under the Training and Mobility of Researchers (FP4) Programme, which was undertaken at the University of Hertfordshire. We thank the U.K. Environment Agency for its support of the work through a small grant from the National Coarse Fish Centre.
References
[1] Nunn A.D., Harvey J.P., Cowx I.G., Variations in the spawning periodicity of eight fish species in three English lowland rivers over a 6 year period, inferred from 0+ year fish length distributions, J. Fish Biol., 2007, 70, 1254–1267
[2] Nunn A.D., Harvey J.P., Britton J.R., Frear P.A., Cowx I.G., Fish, climate and the Gulf Stream: the influence of abiotic factors on the recruitment success of cyprinid fishes in lowland rivers, Freshwat. Biol., 2007, 52, 1576–1586
[3] Britton J.R., Pegg J., Ecology of European barbel Barbus barbus: implications for river, fishery, and conservation management, Rev. Fish. Sci., 2011, 19, 321–330
[4] Peňáz M., Pivnička K., Baruš V., Prokeš M., Temporal changes in the abundance of barbel,
Barbus barbus in the Jihlava River, Czech Republic, Folia Zool., 2003, 52, 441–448
[5] Wheeler A., Jordan D.R., The status of the barbel, Barbus barbus (L.) (Teleostei, cyprinidae), in the United Kingdom, J. Fish Biol., 1990, 37, 393–399
[6] Pilcher M., Copp G.H., Winter distribution and habitat use of fish in a regulated lowland river system of South-East England, Fish. Manage. Ecol., 1997, 4, 199–215
[7] Watkins M.S., Doherty S., Copp G.H., Microhabitat use by 0+ and older fishes in a small English chalk stream, J. Fish Biol., 1997, 50, 1010–1024
[8] Pegg J., Britton J.R., Effects of inter- and intraspecific competition on the growth rates of juvenile European barbel Barbus barbus used in the stock enhancement of UK fisheries, Fish. Res., 2011, 112, 8–12
[9] Copp G.H., Doherty S., Faulkner H., Watkins M.S., Majecki J., Diel drift behaviour of fish eggs and larvae, in particular barbel, Barbus barbus (L.), in an English chalk stream, Fish. Manage. Ecol., 2002, 9, 95–104
[10] Copp G.H., Spathari S., Turmel M., Consistency of diel behaviour and interactions of stream fishes and invertebrates during summer, River Res. Appl., 2005, 21, 75–90
[11] Copp G.H., Behaviour of young and small fishes in the River Lee over a 24-h cycle in mid summer, Lond. Nat., 2004, 83, 87–95
[12] Vilizzi L., Copp G.H., Behavioural responses of 0+ barbel in an artificial stream: distribution and velocity use, Anim. Behav., 2001, 61, 645–654
[13] Vilizzi L., Copp G.H., An analysis of 0+ barbel (Barbus barbus) response to discharge fluctuations in a flume, River Res. Appl., 2005, 21, 421–438
[14] Vilizzi L., Copp G.H., Diel densities of young-of-theyear fishes in the River Lee (Hertfordshire) over a 24-h period in late summer, Lond. Nat., 2008, 87, 113–117
[15] Vilizzi L., Copp G.H., Interstitial movement and emergence of barbel, Barbus barbus, free embryos and larvae in a laboratory flume, J. Fish Biol., 2012, 82, 1057–1063
[16] Campana S.E., Neilson J.D., Microstructure of fish otoliths, Can. J. Fish. Aquat. Sci., 1985, 42, 1014–1032
[17] Peňáz M., Embryonic development of the barb, Barbus barbus (Linnaeus, 1758), Zool. Listy, 1973, 22, 363–374
[18] Escot C., Granado-Lorencio C., Morphology of the otoliths of Barbus sclateri (Pisces: Cyprinidae), J. Zool., 1998, 246, 89–94
[19] Campana S.E., Annand M.C., McMillan J.I., Graphical and statistical methods for determining the consistency of age determinations, Trans. Am. Fish. Soc., 1995, 124, 131–138
[20] Hoenig J.M., Morgan M.J., Brown C.A., Analysing differences between two age determination methods by tests of symmetry, Can. J. Fish. Aquat. Sci., 1995, 52, 364–368
[21] Campana S.E., Accuracy, precision and quality control in age determination, including a review of the use and abuse of age validation methods, J. Fish Biol., 2001, 59, 197–242
[22] R Development Core Team, R: A language and environment for statistical computing, R Foundation for Statistical Computing, Vienna, Austria, 2010
[23] Ogle D., FishR vignette – Precision and accuracy in ages, Northland College, Wisconsin, USA, 2011
[24] Anderson A.J., Gorley R.N., Clarke K.R., PERMANOVA+ for PRIMER: Guide to Software and Statistical Methods, PRIMER-E Ltd, Plymouth, UK, 2008
[25] David A.W., Isely J.J., Grimes C.B., Differences between the sagitta, lapillus, and asteriscus in estimating age and growth in juvenile red drum, Sciaenops ocellatus, Fish. Bull., 1994, 92, 509–515
[26] Krupka I., Early development of the barbel Barbus barbus (Linnaeus, 1758), Práce Úst. Ryb. Hydrobiol., 1988, 6, 115–138
[27] Vilizzi L., Age, growth and cohort composition of 0+ carp in the River Murray, Australia, J. Fish Biol., 1988, 52, 997–1013
[28] Smith B.B., Walker K.F., Validation of the ageing of 0+ carp (Cyprinus carpio L.), Mar. Freshwat. Res., 2003, 54, 1005–1008
[29] Durham B.W., Wilde G.R., Validation of daily growth increment formation in the otoliths of juvenile cyprinid fishes from the Brazos River, Texas, N. Am. J. Fish. Manage., 2008, 28, 442–446
[30] Morioka S., Kaunda E., Otolith growth increments in three cyprinid species in Lake Malawi and information of their early growth, In: Weyl O.L.F., Weyl M.V., (Eds.), Proceedings of the Lake Malawi Fisheries Management Symposium, 4th–9th June 2001, Capital Hotel, Lilongwe, Malawi, 2001, 220–226
[31] Vilizzi L., Walker K.F., Jain T., McGlennon D., Tsymbal V., Interpretability and precision of annulus counts for calcified structures in carp, Cyprinus carpio L., Archiv Hydrobiol., 1998, 143, 121–127
[32] Campana S.E., Gagne J.A., Munro J., Otolith microstructure of larval herring (Clupea harengus): image or reality?, Can. J. Fish. Aquat. Sci., 1987, 44, 1922–1929
[33] Taylor A.A.L., Britton J.R., Cowx I.G., Does the stock density of stillwater catch and release fisheries affect the growth performance of introduced cultured barbel?, J. Fish Biol., 2004, 65, 308–313
