Age and Growth of Triplophysa ( Hedinichthys ) yarkandensis (Day, 1877) in the Tarim River in Xinjiang, China

Influenced by water pollution and flow


INTRODUCTION
A ge determination is the basis for studying the biology and population ecology of fish (Yin, 1995). Age can be determined by three methods: Direct observation, statistics and analyses on catch, and structural analyses on calcification (Ye, 2002). Scales, vertebrae, otoliths, opercular, cleithra, actinosts, basioccipital bones, and hyoid bones are usually used for the observation, analysis, and estimation on age (Wu, 1975;Xie et al., 1994). Previous studies have found out that it is not reasonable to choose scales for those old and slow-growing fishes. Their scales are easy to get lost which results in underestimation (Shen et al., 2001).
For different fishes, different age structures should be utilized. High-accurate structure should be taken as the principal material and others as the supporting materials. Therefore, it is necessary and crucial to select the best and the most suitable material for age determination (Polat et al., 2001). Otoliths have been proved to be the most commonly accurate age structure for fish so far. Ring marks in otoliths grow steadily no matter how widely the living environment changes (Sponaugle, 2009), because it is closely correlated with the growth of fish (Casselman, 1990;Campana, 2001;Horn, 2002;Chen et al., 2009). Pannella (1971) found the daily increment on otolith, which laid a solid foundation for the studies on the relationship between otoliths and the growth of fish.
Triplophysa, widely spread in inland waters on plateaus in the middle parts of Asia, is indispensable to the fish fauna on Qinghai-Tibet Plateau (He et al., 2008), having special adaptability to life on plateau (He, 1996). Triplophysa (Hedinichthys) yarkandensis (Day 1877) (Cyprinidformes, Cobitidae, Triplophysa), is an endemic fish species in the Tarim River. Locally known as Goutou (Dog-head). Since 1960s, population structure has been disrupted by human activities, biological invasion, water O n l i n e

F i r s t A r t i c l e
pollution and other factors. This caused the decline in the numbers of endemic fishes (Wang, 1995;Olden et al., 2010;Xiong et al., 2015). T. yarkandensis may become another indigenous endangered fish species in the Tarim River after Aspiorhynchus laticeps and Schizothorax biddulphi (Zhu, 1989;Chen, 2012). Recently, the research of T. yarkandensis mainly focuses on growth and reproduction, culture and toxicity test, genetic diversity and interspecific differences (Zhao, 1989;Zeng and Tang, 2010). As a plateau fish, the T. yarkandensis has a slower growth rate, smaller size and greater salt tolerance, and has specific economic value and unique ecological significance.
In this study, otoliths were chosen for age determination. Relationship between otoliths and body length/age was described. The von Bertalanffy growth model was fitted and its parameters were compared with results of other studies. The potential for population development and characteristics of the T. yarkandensis population growth in the Alar Section of the Tarim River were identified. It served as a foundation for the correct prediction on the development prospect of the population and the reasonable conservation and effective utilization of the fish resources.

Sampling
Periodic sampling was carried out in February, May, August and November from 2018 to 2020, using a variety of gear, including drift gill net (mesh size 2cm), set gill net (mesh size 2cm) and net cage (mesh size 2cm). 940 T. yarkandensis individuals were collected in the Alar Section of the upper of Tarim River (Figs. 1A, B).
All specimens were measured at the site for standardlength (SL, mm), weight (W, g), gonad weight (W G , g), visceral weight (W V , g) and eviscerated weight (W E , g). SL was accurate to 1.0 mm; while W, W G , W V , and W E to 0.01 g. Biological dissection was performed on site. The left lapillus otoliths, and the 6 th to 7 th vertebrae and opercular were selected as age identification materials. All of them were washed, dried, wrapped in filter paper and put in 2-milliliter plastic centrifuge tubes. The specimens which were taken successfully for all three structures were 891 out of 940.

Preparation and examination of age structures
Otoliths Lapillus was embedded in nail enamel and mounted on a glass slide with the convex side pointing upwards. After dried for 12 h, lapillus was ground with 1,000 to 2,000 grit wet sandpaper. Frequent microscopic observation and

O n l i n e F i r s t A r t i c l e
adjustment were made during the grounding process until the core was reached. Then lapillus was polished. When one side was finished, acetone was applied to dissolve the nail enamel. Lapillus was turned over, mounted, dried, ground and polished in the same way (Xiong et al., 2006;Li et al., 2008;Ma, 2011). When the core plane was available, it was observed and shot under anatomical lens Leica EZ4D and Niko80i at 10×/0.3 magnification. And then the radius of lapillus at 100 um scale were measured using Image-Pro Plus 6.0 (Fig. 1C).
Previous studies have demonstrated that otolith size and weight etc. are highly associated with the growth traits. Only the radius of the sectioned lapillus was measured as the otolith of T. yarkandensis was extremely small.

Vertebrae
As the vertebrae of T. yarkandensis were relatively small, it took only 2 to 5 min to boil them in the boiling water. After the boiling, vertebrae were taken out, dried, and soaked in 1% hydrogen peroxide for 24 to 48 h. Then the 6 th and 7 th vertebrae were removed and put under anatomical lens Motic SMZ-168 for observation and taking photos. If not clear, dimethyl could be added as transparent reagent. If still not clearly read, the vertebrae could be adjusted for a better observation angle (Fig. 1C).

Opercular
Both sides of opercular were taken and put in hot water to boil for 1 min. They were cleaned for residual tissue, dried, and soaked in 1% hydrogen peroxide for 24 to 48 h. When all the above had been completed, the opercular were examined and photographed using an anatomical lens, the Motic SMZ-168.

Analysis of the formation cycle of rings
March to May is the peak breeding period of T. yarkandensis. The formation cycle of rings was analyzed through marginal increments (MIR) (Haas and Recksiek, 1995). The equation is as follows: Where R is the radius of the hard structure, R n is the radius from the core to the last ring on the outside, and R n-1 is the radius from the core to the penultimate ring on the outside.
Photos used in the above analysis were taken under microscope Niko80i at 10×/0.3 magnification and measured at 100 um scale using Image-Pro Plus 6.0.

Age validation
Age was validated according to the ring number on the hard structure. In this study, the first ring was considered one year of age from where age increases (Massutí et al., 2000). Samples were assessed for age independently by two readers. If the ages assigned by each reader agreed, then the ages were considered valid. Otherwise, the two readers re-examined the structure together and reached an agreement. In case the second results still varied widely, the sample had to be abandoned .

Relationship between length and weight
In this study, a common equation was applied to analyze the characteristics of growth. Measurements of length and weight between sexes were used for Kolmogorov-Smirnov test and significance analysis (Cazorla and Sidorkewicj, 2009;Ma, 2011). A t-test was used to compare growth parameters b and 3 (Pauly, 1984) to estimate whether T. yarkandensis grow at a constant speed.
The Power-exponential relationship between length and weight was shown in the following equation:

Growth equation
The characteristics of the growth were described by the classic von Bertalanffy growth model: Where L t is the length at age t of individual fish, W t is the weight at age t, L ∞ is the maximum attainable length, W ∞ is the maximum attainable weight, t 0 is the hypothetical age at which length and weight were zero, and k is the curvature of the growth curve.
In order to further predict the growth trend, the growth rate and acceleration equations were obtained by first order and second order derivation of the growth equation. The growth trend of fish was fitted by the measured age data or the retrogradation age data. And the growth rate was slower when it approached an asymptotic value (Fei and Zhang, 1990;Yin, 1995;Pauly, 1981).

Data processing
All statistical analyses were performed using SPSS 16.0 and Origin 8.0.

Frequency distributions of length and weight
The frequency distributions of SL were shown in Figure 2A. The length ranged from 30.0 to 195.0 mm and the mean SL was 96.5±17.0 mm. The predominating SL ranged from 75.0 to 125.0 mm, accounting for 51.81% of the whole specimens. There were 485 males whose length ranging from 65.0 to 162.0 mm, and 412 females ranging The frequency distributions of W were shown in Figure 2B. The weight ranged from 3.40 to 114.04 g and the mean weight was 14.25±7.29 g. The visceral weight ranged from 0.07 to 12.14 g, and the evisceral weight ranged from 1.06 to 101.59 g with the mean weight of 12.77±8.52 g. The predominating weight ranged from 5.00 to 20.00 g, proportion of those above 30.00 g decreasing as their weight increases. The weight of the indeterminate ranged from 12.57 to 85.48 g. In the whole specimens, the weight of male from 3.40 to 50.99 g and female ranged from 3.59 to 114.04 g. Weight between sexes differ significantly in different ages Kolmogorov-Smirnov test, F = 23.362, p < 0.05).

Frequency distributions of age
The vertebrae and opercular were found to be less useful for determining the age of T. yarkandensis, especially for older fish, and were only utilized as an additional aid for identification in cases of otolith loss. 891 were successfully read for ages in the 940 specimens, and the rest 49 were unsuccessfully read (including 6 females and 43 indeterminate). Age frequency distribution of T. yarkandensis was shown in Figure 2C. The youngest individual in the capture population was 1 year of age, the oldest was 10 years, and the preponderant was 2.5 to 6.5 years. The predominant male was 5 years and female were 4 to 6 years of age.

Relationship between otolith radius and body length/age of fish
The relationship between radius of the major axis of otolith and body length/age was analysed (Fig. 4). The fitting equation of otolith radius (OR) and body length was O n l i n e

F i r s t A r t i c l e
Age and Growth of Triplophysa (Hedinichthys) yarkandensis power function with low correlated: OR = 7.9688L 0.1101 (R 2 = 0.0033). While the fitting equation of otolith radius and logarithmic age (A) represented linear relationship: OR = 293.56A+9336.6 (R 2 = 0.0689). Above means otolith radius increases as body length and age increases.

Growth equation
Length data of individual age classes of both sexes of T. yarkandensis were shown in Table I and Figure 5. Mean length between male and female in the same age class did not differ significantly (independent sample t-test, p > 0.05, except 3+ and 6+). In age class 3+ and 6+, mean length of female differed significantly (p < 0.05). Length differences in the same age class were also significant as T. yarkandensis grow older. They were used to fit growth curves.

Growth rate and acceleration
Performing first-order derivative and second-order derivative based on the above length and weight growth equation. The following equations of growth rate and acceleration were obtained (Fig. 6).  Male and female T. yarkandensis had similar variation trends in the growth rate and acceleration of length and weight (Fig. 5-6). Age at inflection point for male (t i ) was 8.24, length and weight of which was 113.5 mm and 19.26 g, while for female (t i ) was 10.39, length and weight of which was 145.2 mm and 36.47 g.

O n l i n e F i r s t A r t i c l e
We found no inflection point on the curves of growth rate and acceleration of length. Growth rate and acceleration were negatively correlated with age increase.
The weight grew sustainably and the trend was flat at the inflection point. While the acceleration of weight increases first and then decreases, representing a slow decline. As for female, arriving at the maximum when the age was two, which was key to the inflection point of weight. As for male, there was no rising trend but a slow drop, leveling off when age was zero.

DISCUSSION
There has always been disputed about the relationship between the size of otolith or scale and the body length of fish (Yin, 1995). By taking measurements on otolith through different axes, we could find that otolith grow all the time but on a non-uniform rate (Fowler, 1990). For old or slow-growing fish, continuous distal deposition of otolith resulted in non-uniform growth of body length (Boehlert, 1985;Reñones et al., 2007;Ma, 2011), which will reduce the accuracy of back-calculation of length (Campana, 1990).
From the growth equation of T. yarkandensis, we found the age at inflection point was relatively old. 10+ were the most numerous in the specimens and around 9+ for male and female. The relationship between coefficient k and asymptotic length and asymptotic weight was positive, suggesting that T. yarkandensis is a fast-growing, longlived Triplophysa although the growth is slow compared to other fish (Table II). The maximum length and weight of female were bigger than those of male, and the k value of female was smaller than that of male. This was closely related to the big change in the aquatic environment of the Tarim River and the different maturity of individual male and female.
The finding of longevity of T. yarkandensis was similar to that of Zhang et al. (2010). The reasons for different growth parameters obtained by different authors might be the following (Table II): (1) different distributions of length and weight as a consequence of different mesh sizes and number of specimens; (2) different sampling locations, areas and seasons; (3) different methods of fitting growth equations (mainly by back-calculation and field measurement of length). And studies of Cailliet and Goldman (2004) demonstrated that if there are a lack of sufficient number of samples of younger fish or older fish to apply the established model, t 0 will be instability. This can lead to inaccurate inferred growth models. The growth coefficient k was useful during the assessment of the potential sensitivity of fish resources (Musick, 1999;Li et al., 2009). This study showed that the k value of male was higher than that of female. Those slow-growing, long-lived, late-maturing fish, such as T. yarkandensis, were very sensitive to the environment. High mortality and fast consumption of resources make the recruitment rate slower than expected (Musick, 1999). These characteristics were probably due to the adaptation to extreme environments of low water temperature and lack of food.

CONCLUSION
In this study, T. yarkandensis of Tarim River was slow-growing, long lived and the individual was relatively small-sized. The fullness was asynchronous and uneven and the population structure was not stable. They cause the fish resource was limited. For those fish, it is very difficult to recover when the stock is depleted. In conclusion, as an endemic fish, T. yarkandensis deserves our attention and proper conservation. It was quite necessary to set up a set of scientific fisheries management measures for the conservation of T. yarkandensis.

O n l i n e F i r s t A r t i c l e
Age and Growth of Triplophysa (Hedinichthys) yarkandensis O n l i n e