The migratory catfish Calophysus macropterus (Lichtenstein, 1819) is largely distributed in rivers and floodplains of the Amazon and Orinoco basins (Pérez, Fabré, 2009). Being a single species genus in the family Pimelodidae, it was traditionally not a prime target in Amazonian fisheries (Agudelo et al., 2000; Pérez, Fabré, 2002), except in the Putumayo River in Colombia where it has long been one of the main landed species (Bonilla-Castillo et al., 2011). In the last two decades, however, landings of C. macropterus have greatly increased in Colombia and Peru, possibly owing to the decreasing abundance of the main catfish targets such as the large Brachyplatystoma Bleeker, 1862 and Pseudoplatystoma Bleeker, 1862 species and to the growing demand in Colombian big cities, where it is sold fraudulently as the ‘capaz’, Pimelodus grosskopfii Steindachner, 1879, a highly appreciated and now endangered catfish species from the Magdalena River (García Vasquez et al., 2009, García-Vasquez et al., 2012; Bonilla-Castillo et al., 2011, 2012; Mosquera-Guerra et al., 2015). While its catches barely accounted for 0.3% of total landings in the Loreto region (Peru) in 2000, C. macropterus was the 7th most landed species in 2016, accountin`g for ~ 4% of total landings in the Loreto and > 6% in the Ucayali region (García Vasquez et al., unpublished data). In the Putumayo River, Colombia, it already accounted for ~ 21% of total landings in 2000 (Bonilla-Castillo et al., 2012) and in 2016, it accounted for ~ 50% of total landing, together with Pinirampus pirinampu (Spix & Agassiz, 1829), another pimelodid species (Bonilla-Castillo, Agudelo, unpublished data).

Although its diet can be classified as opportunistic omnivorous to piscivorous, based on its main food items: fish (dominant item), plant material (including seeds) and invertebrates (García Vasquez et al., 2017), the species also have a pronounced necrophagous tendency, which explains why it was traditionally rejected in central Amazonia (Pérez, Fabré 2009; Cunha et al., 2015). Additionally, the feeding habits of C. macropterus often result in methyl-mercury bioaccumulation largely exceeding the maximum reference value (0.5 mg.kg^-1) established by the World Health Organisation (Mosquera-Guerra et al., 2015). As a consequence, its fishing has been banned for 5 years (2016-2020) by normative actions in Brazil (Franco et al., 2016) and indefinitely in Colombia since 2017 (resolution Nº1710/2017, AUNAP). In Brazil, using mtDNA sequencing, Cunha et al. (2015) evidenced that C. macropterus processed as fillet was frequently sold by fish-processing plants and markets under deceptive names, in spite to the rejection of scavenger fishes by Brazilian consumers. Although fishing C. macropterus has been officially banned in Brazil, several factors suggest that its fisheries might still be operating: the existence of an established production chain that largely exports to Colombia (Franco et al., 2016) together with a widespread retail market in Brazil (Cunha et al., 2015), the fact that it yields important income with much less effort than other fisheries in the Amazon (Iriarte, Marmontel, 2013) and the lack of enforcement by government agencies (Cunha et al., 2015). Yet, in spite of its recently acquired economic importance and popularity, albeit infamous, in most Amazonian countries, the few studies about the life history characteristics of C. macropterus are geographically restricted to Central Amazonia (Pérez, Fabré, 2002, 2009) and a portion of the Colombian Amazon (Agudelo et al., 2000). Apart from some information about its reproduction in the Putumayo River (Camacho et al., 2006), little is known about its life history traits and population dynamics in the Putumayo and in the Peruvian Amazon, where this species strongly contributes to fisheries landings. The present study aimed at characterising important life history traits (reproductive, growth and mortality patterns) for fisheries management and conservation in the main landing site of the Peruvian Amazon, Iquitos, and in Puerto Leguízamo, located in the upper Putumayo River, shared by Colombia and Perú. Indeed, previous studies have shown important differences in the life history traits of different species between the two river systems (Duponchelle et al., 2012, 2015; Bonilla-Castillo et al., 2018).

Sampling area. The city of Puerto Leguízamo (00°11’53.2”S 74°46’42.7”W) is located in the upper part of the Putumayo River at a height of 220 m.a.s.l. its climate is hot and humid (Bonilla-Castillo et al., 2018) (Fig. 1). Iquitos (03°44’56.83”S 73°15’13.79”W), on the other hand, is located at ~ 90 m.a.s.l., slightly below where the Marañón and Ucayali, two of the major tributaries of the Amazon basin, meet to form the Amazonas River. We will thereafter refer to the lower Marañón, lower Ucayali and Amazonas around Iquitos as the AMU system. The AMU is mainly constituted of nutrient-rich white waters originating from the Andes (Hanek, 1982). The hydrological conditions in the AMU produce flood pulses that inundate large, highly productive areas responsible for the best fisheries production of the Peruvian Amazon (Tello, Bayley, 2001). The Marañón River has a length of 1,707 km, its average velocity on the lower portion is 1.65 m.s^-1 with flow ranging between 7,000 to 25,000 m³.s-1 (MINAM, 2011). The Ucayali River has a length of ,2670 km, its average velocity is between 1.5 to 2.5 m.s^-1, with a flow ranging from 2,700 to 20,000 m³.s^-1 in the lower portion (MINAM, 2011). The Putumayo is one of the two major affluents of Colombia flowing into the Amazon River. It has a length of 2,000 km, of which 1,500 travels the Colombian territory and 450 km in Brazil (Alonso et al., 2006). The Putumayo River is a white-water system of Andean origin. At the Puerto Leguízamo is has a flow of 268 m³.s^-1 average, with a maximum value of 417 m³.s^-1 in the month of May. The transparency of the water is approximately 16 cm, the conductivity of the river is 36.96 μS/cm, pH of 6.26 average, dissolved oxygen of 7.85 mg/L and average water temperature of 27.2ºC (Núñez-Avellaneda et al., 2006).

FIGURE 1

Fish sampling. In the Putumayo, fish were sampled monthly between 2013 and 2017 in the main landing sites of the city of Puerto Leguízamo (Colombia), where fishermen exert their activity ~140 km around the city (Fig. 1). In Iquitos (Peru), fish were sampled monthly between 2013 and 2015 in the main landing sites of the city, where fishermen land their products from the Amazonas, the lower Marañón and the lower Ucayali rivers (AMU system).

In both sampling sites, fishermen mainly use two methods for the capture of C. macropterus. The first is using bait, usually blood and decomposing meat (mainly beef and pork viscera and grease) that is submerged near the boat along the river preferably during the hours of night and early morning. Once the fish have concentrated around the bait, they are caught by purse seines of 2 inches mesh-size. This method is used during all hydrological periods, although less frequently on nights with full moon. It is non selective, as all sizes get caught.

The second fishing method uses a nylon line with 30 to 50 hooks size #6, submerged perpendicularly to the main river channel. Hooks are baited with fish parts (see García Vásquez et al., 2017 for further details). This method is more selective as hook size determines bait size and hence the size of fish caught. But as the same size of hooks is used in both the AMU and the Putumayo, selectivity should not impact the size distribution of fish caught in the two sampling areas.

Specimens were compared with reference collections of this species in the Amazon Icthyological Collection of Instituto Amazónico de Investigaciones Científicas Sinchi (e.g., voucher Sinchi: CIACOL:455, Leticia, 04º12’30.5”S 69º56’0.5”W, Amazon River, Amazonas, Colombia) to verify identification.

For each fish, standard length (SL) was recorded to the nearest cm, and when possible total and eviscerated body mass (to the nearest g, Wt and We, respectively). When the fish were not eviscerated, gonads were photographed and macroscopically checked for gender and maturity stage determination, then weighed to the nearest 0.1 g. The gonadal maturation scale used was that of Núñez, Duponchelle (2009). Briefly, males at stage 1 are immature, at stage 2 they are either maturing or resting and at stage 3 they are ripe. Females at stage 1 are also immature; at stage 2 they are maturing; at stage 3 they are in advanced maturation; at stage 4 they are ripe; at stage 5 they have just spawned and at stage 6 they are resting.

Reproductive aspect. The reproduction period was estimated from the monthly proportions of gonadal maturation stages and of the gonado-somatic index (GSI = gonad mass/total body mass *100) in adult females. The average size at first sexual maturity (L₅₀) is defined as the standard length at which 50% of the individuals are at an advanced stage of the first sexual cycle during the breeding season (Sparre, Venema, 1998). Practically, this is the size at which 50% of the fish have reached stage 2 of the maturity scale for females and males (Núñez, Duponchelle, 2009). For each river system and for each gender, the mean S_L at first maturity (L₅₀) was estimated by fitting the fraction of mature individuals per 5 mm S_L intervals to a logistic regression function weighted by the total number of individuals in each size class (Bonilla-Castillo et al., 2018):

% M = 100 / 1 + e - a x L - L 50

where %M = percentage of mature individual by size class, L = central value of each size class, and a and L₅₀ = constants of the model.

Growth aspect. Age and growth characteristics were calculated from the modal progressions of SL frequency distributions (King, 2007) using the ELEFAN (Electronic Length Frequency Analysis) routine (Pauly, David, 1981) provided in the FiSAT II (FAO-ICLARM Fish Stock Assessment Tools) package (http://www.fao.org/fi/statist/fisoft/fisat/index.htm) (Gayanilo et al., 2005). The set of parameters that best corresponded to the breeding patterns observed for the species (i.e., which gave an estimated birth date corresponding to the breeding peak) and that best described the distributions (i.e., which went through the largest number of large modes and yielded the largest Score = “goodness-of-fit” parameter of the ELEFAN routine) was selected (García Vásquez et al., 2009). This process also permitted to diminish the tendency of ELEFAN method to underestimate K and overestimate L_∞ (Sparre, Venema, 1998). The growth parameters were calculated by the von Bertalanffy Growth Function (VBGF) equation fitted by the ELEFAN method:

L t = L ∞ 1 - e - K t - t 0

where L_t is the mean length at age t, L_∞ is the asymptotic length, K the growth coefficient and t₀ the theoretical age at size 0.

t₀ was calculated using the empirical formula proposed by Pauly (1980):

l o g 10 t 0 = - 0.392 - 0.275   l o g 10 L ∞ - 1.038   l o g 10 K

The estimated L_∞ and K values were used to calculate the growth performance index (phi prime, ǿ) defined by Pauly, Munro (1984) as

ǿ = l o g K + 2 log ⁡ L ∞

The age at first sexual maturity (A₅₀) was calculated from the VBGF as follows (Duponchelle et al., 2007):

A 50 = - l n 1 - L 50 . L ∞ - 1 K - 1 + t 0

where L₅₀ is the size at first sexual maturity and L_∞ and K are parameters from the VBGF.

The longevity (t_max) was estimated as the age at 95% of L_∞ from the equation of Taylor (1958):

A p = t 0 - l n 1 - p K - 1

where t₀, and K are the VBGF parameters and p is a fraction of L_∞ (in this case 0.95). The longevity was also calculated from the equation of Froese, Binohlan (2000):

log₁₀ t_max = 0.5496 + 0.957 log₁₀ (A₅₀) where A₅₀ is the age at first sexual maturity.

Mortality. Mortality parameters were also estimated using procedures provided in the FISAT II package. Total mortality (Z) was estimated by the method of the length-converted catch curves (LCCC) (Pauly, 1983). Natural mortality (M) is one of the most complicated life history parameters to estimate in natural populations (Pauly, 1980) and several empirical models linking M to life history attributes such as age at maturity or growth were proposed for fish (reviews in Gislason et al., 2010). These empirical relationships assume that M is a species- or stock-specific constant, and users generally apply the estimate to all exploited ages and sizes of the species or stock under study. Several of these relationships were calculated in order to provide a range of natural mortality values and assess their effect on the estimation of fishing mortality and exploitation rate (E). Hence, natural mortality was evaluated using Pauly (1980)’s equation, as implemented in the FiSAT package, for a mean annual temperature of 27°C (Bonilla-Castillo et al., 2018):

- 0.006 - 0.270 * L o g 10 L ∞ + 0 . 6543 * L o g K + 0.4634 * L o g 10 T 0 ,

where L_∞ and K are the VBGF parameters and T° the mean annual temperature.

Other empirical methods were used to estimate natural mortality (M), among them: Equation of Gislason et al. (2010) with the mathematical expression 0.55   -   1.61 L n   ( L )   +   1.44 L n   ( L ∞ )   +   L n   ( K ) ; Rikhter, Efanov (1976): 1.521 / L 50 0.72 - 0.155 ; Charnov et al. (2013): L / L ∞ - 1.5 × K ; Jensen (1996) using K :   1.6 * K ; Jensen (1996) using A 50 : 1.65 / A 50 (A₅₀: Age at which 50% of the population reaches sexual maturity.

For the equations of Gislason et al. (2010) and Charnov et al. (2013) “L” was established as the average length in immature and for mature individuals.

Fishing mortality (F) was calculated as F   =   Z - M (Pauly, 1980). The exploitation rate was estimated as E = F . Z - 1 . For comparison purposes, both F and E were calculated for the different estimates of M.

Biometric indicator. The weight-length relationship was estimated by expression W t = a * S L b , after logarithmic transformation l o g W t = l o g a + b * l o g S L , Wt the total weight, SL the standard length, “a” the intercept of the curve and “b” the allometry coefficient of the weight-length ratio (Zar, 2010).

Statistical analyses. Differences in mean Ls, Wt and We between sexes were tested with a Mann-Whitney rank-sum test. Parametric tests could be carried out as either normality of distributions or homocedasticity were not met. Differences in size at first sexual maturity were tested using t-test.

A total of 8,064 specimens of C. macropterus were analysed from Puerto Leguízamo (upper Putumayo River), 3.2% (254) in 2013 (November and December), 26.9% (2,161) in 2014, 28.9% (2,319) in 2015, 31.1% (2,528) in 2016 and 10% (802) in 2017 (January to June). A total of 1,710 specimens were analysed from Iquitos, 37% (634) in 2013, 40% (688) in 2014 and 23% (388) in 2015. The smallest size class observed in the Putumayo River was 22 cm, whereas individuals of 16 cm were caught in commercial fisheries of the AMU system around Iquitos (Figs. 2A, B). Overall, in both river systems females were significantly larger (Mann-Whitney Rank Sum test: U = 401256, P<0.001 in the Putumayo River; U = 18224, P<0.001 in the AMU rivers) and heavier (U = 466462.5, P<0.001 in the Putumayo River; U = 17729.5, P<0.001 in the AMU) than males in commercial landings (Tab. 1).

FIGURE 2

TABLE 1

River	Sex	N	SL mean (cm)	(SL cm) range	Wt Mean (g)	Wt Range (g)	a (CI95%)	b (CI95%)	r2
Upper Putumayo	F	2706	39.0	23.0-47.5	807	128-1453	0.010 (0.009-0.011)	3.08 (3.04-3.12)	0.9014
	M	4115	31.0	22.0-45.8	395	140-1254	0.014 (0.012-0.015)	2.99 (2.95-3.02)	0.8898
	T	8064	34.0	22.0-47.5	550	128-1454	0.011 (0.010-0.011)	3.06 (3.04-3.07)	0.9578
AMU	F	1112	35.0	23.0-46.6	588	61-1312	0.009 (0.008-0.011)	3.08 (3.04-3.12)	0.9656
	M	597	28.4	22.0-45.8	304	74-1020	0.011 (0.009-0.013)	3.04 (2.98-3.09)	0.9510
	T	1710	32.7	16.5-46.6	489	61-1312	0.010 (0.009-0.010)	3.08 (3.07-3.10)	0.9623

Although maximum observed length (Lmax) only varied by 4% and 2% between females and males in the Putumayo and AMU, respectively, maximum observed body mass (Wmax) varied by 14% and 22% in the Putumayo and AMU, respectively. However, when looking at mean observed lengths and masses, differences between males and females varied by 20% in length and 51% in mass for the Putumayo and by 19% in length and 48% in mass in the AMU (Figs. 2A, B; Tab. 1).

Over the study period, standard length (SL) - total mass relationships (M_t) were Wt = 0.0099 SL^3.078 (R² = 0.901, P<0.001) and Wt = 0.0135 SL^2.988 (R² = 0.890, P<0.001) in the Putumayo for females and males, respectively and Wt = 0.0093SL^3.081 (R² = 0.966, P<0.001) and Wt = 0.0109 SL^3.037 (R² = 0.951, P<0.001) in AMU for females and males, respectively.

Reproduction. In the upper Putumayo River, both the gonado-somatic index (GSI) and the monthly proportions of ripe females indicated a breeding period between December and March, corresponding to the low water and early flooding period (Fig. 3A). This pattern was consistent across years. On the other hand, not a single ripe female or even a female with developed gonads was observed in the AMU (Iquitos area) during the two and a half years of study. This is also reflected in the constantly very low GSI values (Fig. 3B).

FIGURE 3

Age and growth. The S_L frequency histograms and the corresponding VBGF curves for the whole time series are presented for all individuals (sexed and unsexed pooled together). The best fitting models obtained from the ELEFAN routine gave mean birth dates corresponding to the peak of the breeding season in the upper Putumayo River (Fig. 4A). As no single reproductively active individuals could be observed in the AMU around Iquitos during the period of study, the best fitting model giving a mean birth date during the low-water season, beginning of the rising water period (corresponding to the observed breeding period in the upper Putumayo) was selected (Fig. 4B). The corresponding VBGF parameters for males, females, and all individuals (sexed and unsexed) pooled together, are presented in Tab. 2.

FIGURE 4

TABLE 2

	Sex	L∞ (cm)	K (Year-1)	t0 (Year-1)	ǿ	Score	tmax1 (Year)	tmax2 (Year)
Upper Putumayo	Females	50.4	0.40	-0.357	3.007	0.146	7.1	5.8
	Males	48.4	0.45	-0.319	3.023	0.150	6.3	5.3
	Total	50.1	0.42	-0.340	3.023	0.141	6.8	5.1
AMU	Females	49.9	0.68	- 0.21	3.229	0.394	4.2	3.9
	Males	43.9	0.80	- 0.35	3.188	0.255	3.6	3.2
	Total	50.0	0.6	- 0.29	3.176	0.171	4.7	3.9

In both the upper Putumayo and in the AMU, females tended to have a better growth than males as indicated by the calculated SL-at-age (Tab. 3). Females were systematically larger than males and the size difference tended to increase with age. This among-gender difference, however, tended to be greater in the AMU than in the upper Putumayo. Fish also tended to have a better growth in the AMU than in the upper Putumayo: this difference ranged from > 21% at age one to 2% at age 8 (see # Total, Tab. 3). The youngest individuals sampled were ~ 1 year old (22 cm male) in the upper Putumayo and ~ 4-month-old (16.5 cm female) in the AMU.

TABLE 3

	Upper Putumayo				AMU					A
Age (t)	F	M	Total	# F-M	F	M	Total	# F-M	# Total	Total
1	21.1	21.7	21.6	- 2.6%	27.9	26.8	26.2	4.0%	21.3%	18.6
2	30.8	31.4	31.3	- 1.9%	38.8	36.2	36.9	7.0%	19.9%	26.8
3	37.2	37.5	37.8	- 0.8%	44.2	40.5	42.8	9.4%	13.2%	32.1
4	41.6	41.5	42.0	0.3%	47.0	42.4	46.1	11.0%	9.8%	35.7
5	44.5	44.0	44.8	1.2%	48.4	43.2	47.8	12.1%	6.7%	38.1
6	46.4	45.6	46.6	1.9%	49.1	43.6	48.8	12.8%	4.7%	39.6
7	47.7	46.6	47.8	2.4%	49.5	43.8	49.3	13.1%	3.1%	40.7
8	48.6	47.3	48.6	2.9%	49.7	43.8	49.6	13.4%	2.1%	41.4

The calculated life span for C. macropterus varied according to the model used (Tab. 2). Froese, Binohlan’s (2000) model resulted in life span estimates of 5.8 and 5.3 years for females and males, respectively, in the upper Putumayo, and of 3.9 and 3.2 years in the AMU. Taylor’s model (1958) gave higher estimations of 7.1 and 6.3 years for females and males, respectively in the upper Putumayo, and of 4.2 and 3.6 years in the AMU. Computing age from the VBGF, the largest fish sampled (47.5 cm unsexed individual) was ~ 7 years old in the upper Putumayo and ~ 4 years old (46.6 cm female) in the AMU (Tab. 3).

Growth and mortality patterns. Over the study period, males and females reached maturity at remarkably similar sizes in the upper Putumayo and AMU (Fig. 5): 28.2 cm (±0.15 SD) and 25.6 cm (± 4.45 SD) for females and males, respectively in the upper Putumayo; 27.9 cm (± 8.55 SD) and 24.6 cm (± 8.15 SD) in the AMU. Consequently, there was no significant difference in size at maaturity between the upper Putumayo and AMU for both females (t = 0.090, df = 28, P = 0.993) and males (t = 0.108, df = 26, P = 0.915). Similarly, among-sex differences were not significant either in the upper Putumayo (t = 0.530, df = 26, P = 0.596) or in the AMU (t = 0.279, df = 28, P = 0.782). Given the observed growth differences, however, age at maturity differed more markedly between the upper Putumayo (1.7 years old for females and 1.4 years old for males) and the AMU (1 year old for females and 0.8-year-old for males).

FIGURE 5

Length converted catch curves resulted in total mortality (Z) estimates of 2.18 year^-1 for females, 2.48 for males and 2.01 for sexed and unsexed individuals combined in the upper Putumayo (Fig. 6) and of 2.81 for males and 2.67 for sexed and unsexed individuals combined in the AMU (Fig. 7). Natural mortality (M) estimates, using both size-independent (assuming a constant M for all age and size classes) and size-dependant models (taking into account two size classes: mean length of immature fish and mean length of mature fish in the population) are presented in Tab. 4 for both the upper Putumayo and the AMU.

FIGURE 6

FIGURE 7

TABLE 4

Formulae used for estimating M		Putumayo			AMU
		M	F	E	M	F	E
Females	Pauly (1980)	0.83	1.35	0.62	1.18	2.12	0.64
	Richter, Efanov (1976)	0.90	1.28	0.59	1.37	1.93	0.58
	Jensen (1996). Using K	0.64	1.54	0.71	1.09	2.21	0.67
	Jensen (1996). Using A50	0.99	1.19	0.55	1.66	1.64	0.50
	Gislason et al. (2010). Immature	1.06	1.12	0.51	1.92	1.38	0.42
	Gislason et al. (2010). Mature	0.54	1.64	0.75	0.99	2.31	0.70
	Charnov et al. (2013). Immature	1.10	1.08	0.49	1.99	1.31	0.40
	Charnov et al. (2013). Mature	0.59	1.59	0.73	1.07	2.23	0.68
Males	Pauly (1980)	0.91	1.57	0.63	1.37	1.45	0.51
	Richter, Efanov (1976)	1.07	1.41	0.57	1.86	0.95	0.34
	Jensen (1996). Using K	0.72	1.76	0.71	1.28	1.53	0.54
	Jensen (1996). Using A50	1.22	1.26	0.51	2.44	0.37	0.13
	Gislason et al. (2010). Immature	1.07	1.41	0.57	2.11	0.70	0.25
	Gislason et al. (2010). Mature	0.82	1.66	0.67	1.35	1.46	0.52
	Charnov et al. (2013). Immature	1.12	1.36	0.55	2.15	0.66	0.23
	Charnov et al. (2013). Mature	0.87	1.61	0.65	1.42	1.39	0.5
Total	Pauly (1980)	0.86	1.15	0.57	1.09	1.58	0.59
	Jensen (1996). Using K	0.67	1.34	0.67	0.96	1.71	0.64
	Gislason et al. (2010). Immature	1.06	0.95	0.47	1.70	0.97	0.36
	Gislason et al. (2010). Mature	0.65	1.36	0.67	0.87	1.80	0.67
	Charnov et al. (2013). Immature	1.11	0.90	0.45	1.76	0.91	0.34
	Charnov et al. (2013). Mature	0.71	1.30	0.65	0.95	1.72	0.65

Natural mortality estimates calculated from Pauly’s (1980) equation, as implemented in FiSAT II, were within the range of values obtained from the other size-independent models for females and males in both the upper Putumayo and the AMU. In the upper Putumayo, size-independent models (Richter and Efanov and Jensen) tended to give M estimates close to those observed for immature fish and superior to those obtained for mature fish using the size-dependent models (Gislason et al., 2010; Charnov et al., 2013). All these natural mortality estimates resulted in high estimates of fishing mortality (F, 1.19 to 1.64 for females and 1.26 to 1.76 for males) and exploitation rate (E, 0.55 to 0.75 for females and 0.51 to 0.71 for males), excluding immature individuals. Immature individuals also had relatively high F and E values.

In the AMU, size-independent models (Richter and Efanov and Jensen) tended to give M estimates intermediate between those observed for immature and mature fish using the size-dependent models (Gislason et al., 2010; Charnov et al., 2013). Excluding immature individuals, all these natural mortality estimates resulted in higher estimates of fishing mortality (F, 1.64 to 2.31) and lower exploitation rate (E, 0.55 to 0.75) than in the Putumayo for females but lower estimates of fishing mortality (F, 0.37 to 1.53) and lower exploitation rate (E, 0.13 to 0.54) for males.

Reproduction. Published information about the reproductive biology of C. macropterus in the Amazon basin is scant and circumstantial. In the Caquetá River around La Pedrera, in Colombia, Agudelo et al. (2000) mentioned the presence of mature individuals in July, during the early decreasing water period, without showing data. Pérez, Fabré (2002, 2009) did not get any ripe female during their sampling in Central Amazonia (around Manaus) and most individuals were either immature or resting. They suggested, from the slightly higher gonado-somatic index values during the early rising waters, that reproduction probably occurred during that period. The same situation was observed in the present study around Iquitos, where no single ripe female or female with gonads in advanced maturation process was observed during the two and half year’s sampling in the AMU, where the breeding cycles of many species (García-Vasquez et al., 2015; Duponchelle et al., 2015), including pimelodid catfishes (García Vasquez et al., 2009) were successfully studied. This situation was also observed in Bolivia around Trinidad, where the breeding cycles of many species (Loubens, 2003), including pimelodid catfishes (Loubens, Panfili, 2000) were successfully studied. The only location in the Amazon basin so far where ripe individuals were officially reported is in the upper Putumayo around Puerto Leguízamo (Camacho et al., 2006, present study). The present study clearly demonstrates, during over three consecutive hydrological cycles, that the breeding season of C. macropterus occurs during the low water and early rising water periods. This is consistent with previous studies on fish larvae, collected using plankton nets and specifically identified by barcoding/metabarcoding (COI gene), which found C. macropterus larvae in the lower Marañón and lower Ucayali during the low water and early rising water periods (García-Dávila et al., 2015; Mariac et al., 2021). But larvae can travel long distances downstream.

The available evidence strongly suggests that environmental requirements for the reproduction of C. macropterus are met higher in the hydrographical networks, closer to the Andes. Although the elevation of the Putumayo in Puerto Leguízamo is only slightly higher than that of the Ucayali in Pucallpa (~177 m vs. ~154 m, respectively), it is much closer to the Andean piedmont and the slope is much steeper upstream than in the Ucayali around Pucallpa. Similar topographical situations are found in 1) Puerto Maldonado in the upper Madre de Dios (Peru), where fishermen report the occurrence of ripe females between August and October (Duponchelle F., 2021, pers. comm.), corresponding to the low water, early rising water period (Cañas, Pine, 2011); 2) Puerto Villaroel in the upper Mamoré River (Bolívia), where ripe individuals are also observed (Duponchelle F., 2021, pers. comm.). Ripe individuals were also observed in the Portugues, Guanare, Bocono, Masparro rivers (tributaries of the Apure River) in the Andean foothills of Venezuela (300-150 m.a.s.l), during the period of rising waters (April to June; Bonilla Castillo C., 2021, pers. comm.).

As the species is distributed in the whole Amazon basin, the available evidence indicating that it breeds only in the Andean foothills above ~150 m.a.s.l. suggest that its life cycle involves extensive migrations between the breeding areas and the distributing areas as far as Central or Lower Amazonia. Migrations at such a large scale are only known within another Pimelodid catfish genus, Brachyplatystoma spp. (Barthem et al., 1997, 2017; see Duponchelle et al., 2021 for review). This hypothesis, amenable to testing using natural tags such as otolith microchemistry (see review in Hermann et al., 2021), illustrate the large knowledge gaps remaining in the ecology of Amazonian species, including economically important fisheries resources such as Colossoma macropomum (Cuvier, 1816).

Despite the important sexual dimorphism in both mean Lmax and Wmax in the Putumayo and in the AMU, L₅₀ were remarkably similar among genders. Size at first sexual maturity did not differ either between the Putumayo and the AMU although age at first sexual maturity did differ, owing mainly to the observed growth differences among river systems.

Growth and mortality patterns. The important size dimorphism observed between females and males is a common feature in pimelodid catfishes (Agudelo et al., 2000; Loubens, Panfili, 2000; García Vasquez et al., 2009) and characids (Duponchelle et al., 2007), and was also reported for C. macropterus in Bolivia (Loubens, Aquim, 1986). Yet, Loubens, Aquim (1986) observed that sexual dimorphism was particularly pronounced in this species, reaching 20% in maximum observed length (Lmax) and 50% in maximum observed body mass (Wmax), which is consistent with our own observations. Lmax in this study (47.5 and 46.6 cm in the Putumayo and AMU, respectively) were smaller than those reported almost 20 years ago by Agudelo et al. (2000) in the same rivers (52 cm in the Putumayo, 52 cm in the Amazonas close to Leticia, Colombia). Agudelo et al. (2012) also reported this Lmax reduction of C. macropterus in the Putumayo River and suggested it was likely due to the increased exploitation.

In both the upper Putumayo and the AMU, C. macropterus presents a fast growth during the first two years, as was also reported in Central Amazonia (Pérez, Fabré, 2009). In the two river systems, females had a faster growth than males and the difference increased with age, which was not observed in Pérez and Fabré (2009), where both sexes had similar growths. Growth dimorphism was reported in other species of the family Pimelodidae such as Brachyplatystoma rousseauxii (Castelnau, 1855) (García Vásquez et al., 2009) or Pseudoplatystoma tigrinum (Valenciennes, 1840) and P. fasciatum (Linnaeus, 1766) (Loubens, Panfili, 2000). The among-sex difference, however, was much greater in the AMU than in the upper Putumayo. Fish also had a better growth in the AMU than in the upper Putumayo although this difference decreased with age from > 21% at age one to 2% at age 8. The growth rate observed in the present study were both better than that observed in Central Amazonia (Pérez, Fabré, 2009). The growth rate observed in the present study, however, remains low compared to growth rates obtained in captivity, where well fed and cared for specimens can reach in four to five months the length observed at one year old in the Putumayo and AMU, respectively (Kossowski, 1998). The length calculated for one-year-old fish by Pérez, Fabré (2009; Tab. 3 present study) was reached after only two months in captivity (Kossowski, 1998).

In both river systems, females tended to grow larger and faster than males, and the growth difference between females and males tended to be more important in the AMU than in the Putumayo. While we cannot rule out the possibility that the AMU would be a feeding site, while the Putumayo works as a spawning ground, the observed differences could also result from better trophic conditions in the AMU than in the Putumayo. When trophic conditions are less favourable, females cannot find enough energy to express their better growth potential to the maximum, which has already been observed for the red piranha, Pygocentrus nattereri Kner, 1858 (Duponchelle et al., 2007). Although there is no data providing direct evidence of less favorable conditions in the Putumayo, this hypothesis finds support in the fact that and fish generally tend to be smaller in the Putumayo than in the Caquetá (ABC, pers. obs.) and that slower growth were already reported for other species in this river: Prochilodus nigricans Spix, Agassiz, 1829 (Bonilla-Castillo et al., 2018), Osteoglossum bicirrhosum (Cuvier, 1829) (Duponchelle et al., 2015), Pseudoplatystoma punctifer (Castelnau, 1855) (Armas et al., in press).

As previously emphasized (Bonilla-Castillo et al., 2018), however, inter-basin differences in fishery exploitation could also account for slower growth in fish populations of the Putumayo River, where the more heavily exploited populations tend to have smaller maximum sizes and slower growth. This phenomenon is known as tropicalization of stocks (Stergiou, 2002). As natural mortality (M) is one of the most complicated life history parameters to estimate in natural fish populations (Pauly, 1980), estimations from several different methods were calculated in order to provide a measure of reliability. Natural mortality estimates calculated from Pauly’s (1980) equation, as implemented in FiSAT II, were within the range of values obtained from the other size-independent models for females and males in both the upper Putumayo and the AMU. Whatever the method used for calculation, fishing mortality and exploitation rates were high for both the Putumayo and the AMU, indicating overexploitation of C. macropterus in both river systems. Fishery exploitation can therefore not be ruled out in the explanation of the observed slower growth in the Putumayo, although further studies will be needed to sort out its importance relative to the hypothesis of poorer trophic conditions in the Putumayo.

Calophysus macropterus has become quite infamous in most Amazonian countries in recent years because of the conservation and health issues associated with its exploitation and consumption. Yet, strong evidence suggests that in absence of proper enforcement, which is most challenging in the Amazon basin, its exploitation and fraudulent commercialization (i.e., Cunha et al., 2015) will likely continue, further endangering the natural populations of aquatic mammals and reptiles. Given its excellent potential for aquaculture (see Kossowski, 1998 for further details) and its relatively low trophic level (Agudelo, 2015), a complementary way of lowering its impact on natural populations of river dolphins, manatees and caimans, while providing healthy animal protein for the growing human populations, would be to encourage the development of its aquaculture in Amazonian countries, using local products as alternative to fish meal and oil. This solution would also allow relieving some of the harvesting pressure on natural populations of C. macropterus.

Life history traits of Calophysus macropterus varied importantly between the Putumayo and the AMU systems. Reproduction was only observed in the upper Putumayo, suggesting the species needs specific environmental conditions found close to the Andean piedmont and not available in our AMU sampling locations. Females tended to have a better growth than males in both river systems and fish of the AMU tended to grow faster in the AMU than in the Putumayo. The mortality data indicate overexploitation of the species in both river systems.