Aquaculture Research, 2011, 42, 815^822 doi:10.1111/j.1365-2109.2011.02886.x

Reproductive success and fry production of the paiche or pirarucu, gigas (Schinz), in the region of Iquitos, Peru´

1,2 3 1 3 3 Jesu¤ sNu¤ nì ez , Fred Chu-Koo , Magali Berland , Lamberto Are¤ valo , Ola¡ Ribeyro , Fabrice Duponchelle1,2 & Jean Franc° ois Renno1,3 1IRD UR 175 CAVIAR, Montpellier, France 2UNFV-FOPCA, Mira£ores, Lima, Peru¤ 3Instituto de Investigaciones de la Amazon|¤a Peruana ^ IIAP,Programa para el Uso y Conservacio¤ n del Agua y sus Recursos (AQUAREC), Iquitos, Peru¤

Correspondence: JNu¤ nì ez, 351RUE J.F. BRETON, BP 5095, Montpellier cedex 05, France. E-mail: [email protected]

Abstract Keywords: arapaima, pirarucu, paiche, reproduc- tion, peru, ¢sh culture (paiche) is the largest scaled ¢sh living in the . Its biology is both fascinating and misunderstood. In a context of over¢shing, hence reduced natural populations, Introduction of a ¢sh with such interesting charac- teristics (large size, high growth rate, no intramuscu- Arapaima gigas (Schinz) is the largest scaled ¢sh lar spines) is an important issue. The development species living in the Amazon basin (Chu-Koo & of farming production would also reduce the ¢shing AlcaŁ ntara, 2009). It is an emblematic species of the pressure on natural populations and allow re- Amazonian ¢sh fauna, also known as the Paiche in stocking programmes in certain areas. To determine Peru and Pirarucu in Brazil. This species is very at- what factors may in£uence the reproductive success tractive for aquaculture in theAmazon region, owing in captivity, data from breeding reports for 2007^ to its many advantages. Arapaima gigas has the best 2010 were collected among ¢sh farmers in the region growth rate among Amazonian cultivated ¢sh spe- of Iquitos. In parallel, we carried out physico- cies, of 10^15kg per year (Bard & Imbiriba, 1986; chemical measurements in di¡erent ponds where Pereira Filho & Roubach, 2005; Nunì ez, 2009; Rebaza, these paiches breed, and conducted personal inter- Rebaza & Deza 2010), and its ¢let, which is very views about the general ¢sh management condi- popular, does not have intramuscular spines. It also tions. The results show that reproduction occurs accepts low water dissolved oxygen levels, due to its throughout the year but with a higher intensity obligatory aerial breathing. Finally, it has a carnivor- during the rainy season. It also highlights farms ous diet but is still £exible because it can accommo- that have performed much better than others, but date consumption of dead ¢sh or ¢sh feed pellets. no single factor except feeding level has been clearly However, natural populations have been subjected to associated with reproductive success. The environ- excessive ¢shing in the past decades for human con- mental control of reproduction in paiche, therefore, sumption and ornamental ¢sh trading (Chu-Koo & remains partly mysterious. To deepen this study, we Tello, 2010), and already for more than 15 years, recommend the systematic sexing of breeders, ex- A. gigas has become a very uncommon species in the tending reproductive behavioural studies, and exam- Amazon, except in some natural reserves (Pacaya ining the limnological factors involved in fry Samiria in Peru or MamirauaŁ in Brazil). The Paiche mortality. is now in the CITES II list, which means that its trade

r 2011 Blackwell Publishing Ltd 815 Reproductive aspects of ARAPAIMA in captivity JNu¤ nì ez et al. Aquaculture Research, 2011, 42, 815^822 and use are strictly controlled. Scarce available infor- inhibit A. gigas reproduction in controlled environ- mation concerns mainly the reproductive biology ments. in the wild (Guerra, 1980; Saavedra Rojas, Quintero Pinto & Landines Parra 2005); however, it is in the interest of ¢sh farmers to develop a controlled pro- Material and methods duction of this species (Pereira Filho & Roubach, origin 2005; Nunì ez, 2009) and one of the collateral ex- pected e¡ects of Paiche domestication is a decreased Between 2000 and 2007, a programme to assist ¢sh ¢shing pressure on wild populations. Nevertheless, farmers in the region of Iquitos enabled the Instituto many attempts have been made in recent decades to de Investigaciones de la Amazon|¤a Peruna (IIAP) to develop its production in captivity (Bard & Imbiriba, provide six paiche breeders to selected local ¢sh 1986; Imbiriba,1991; Alcantara & Guerra,1992; Guer- farmers. Arapaima breeders had been bred previously ra, AlcaŁ ntara, Padilla, Rebaza,Tello, Isminì o, Rebaza, in ponds for 3 years in the IIAP’s facilities of Quisto- Deza, Ascon, Iberico, Montreuil & Limachi 2002; cocha. Pereira Filho & Roubach, 2005; Saavedra Rojas, Quin- Fish reproduction supervising, mandated by law, tero Pinto, Lopez Hernandez & Pezzato 2005), but this started in 2007 and ¢sh farmers had to report to local task is made di⁄cult by the dispersed knowledge on authorities all events (births, fry harvest, etc.)that oc- A. gigas biology in captivity, especially in crucial as- curred in their farms, thus facilitating data collec- pects like reproduction and feeding. tion. Some farms already owned paiches before the Paiche domestication faces the di⁄culty of gender assistance programme; they did not originate from determination, which is virtually impossible from IIAP’s stock but contributed to some reproductive external characteristics except during the reproduc- events. Arapaima breeder weights varied approxi- tive period; otherwise, sex has to be determined in mately from 60 to100kg. blood or plasma samples by biochemical methods using vitellogenin or sex steroid ratios (Chu-koo, Dugue¤ ,AlvaŁ n Aguilar, Casanova Daza, AlcaŁ ntara Fish rearing and feeding Bocanegra, ChaŁ vez Veintemilla, Duponchelle, Renno, The paiches are fed live ¢sh (forage ¢sh), dead ¢sh Tello & Nunì ez 2009). Moreover, the ovary is com- (trash ¢sh from local market) and sometimes down- posed only by a single naked lobe or gymnovarium graded chicken or chicks, ¢sh or chicken viscera, ¢sh (Godinho, Santos, Formagio & Guimaraes-cruz feed pellets and even fruits, although ¢sh-based food 2005; Saavedra Rojas, Quintero Pinto & Landines is predominant in every farm.The amount of food per Parra 2005; Nu¤ nì ez & Duponchelle, 2009) that makes 1 capita (calculated as kg food paiche biomass À )var- the practice of canulation to determine oocyte devel- ied greatly among farms, and most farmers could not opment, induction of oocyte maturation and arti¢cial precisely quantify how much food they had distribu- fertilization almost impossible. As the production of ted to ¢sh daily, monthly or even yearly. For this rea- ¢ngerlings actually relies only on spontaneous re- son, we had to split the farms in three feeding levels:1, productions in ponds, the knowledge of the optimal small quantity (the forage ¢sh are distributed about requirements for reproductive success is a prerequi- once a month); 2, average quantity (the ¢sh are fed site to substantially increase the ¢ngerling o¡er in about once a week) and 3, large quantity (the ¢sh are this region. Nevertheless, a noticeable development fed daily). In the ¢rst level, the food is essentially for- of A. gigas production is expected as most of the age ¢sh consumed progressively by paiches. In the actual ¢ngerling production is destined for aquacul- other two levels, the food is mainlycomposed of trash ture instead of ornamental ¢sh trade as before ¢sh supplemented, or not, with the other cited food- (Chu-Koo & Tello, 2010). This study aims at analysing stu¡s. the reproductive data collected from A. gigas farmers in the region of Iquitos between 2007 and 2010. The results reported here represent the ¢rst attempt Physicochemical parameters to provide answers on the magnitude of a sustained production level of ¢ngerlings with the current The physicochemical characteristics of pond waters knowledge and management conditions. It also were measured within a weekly period for all farms evaluates the variations of some physicochemical during the rainy season (February^March), using and environmental factors that could promote or the Lamotte AQ-2 kit (Drop Count Titration, Chester-

816 r 2011 Blackwell Publishing Ltd, Aquaculture Research, 42,815^822 Aquaculture Research, 2011, 42, 815^822 Reproductive aspects of ARAPAIMA in captivity JNu¤ nì ez et al. town, MD, USA). The following parameters were composed of a few ponds of di¡erent sizes and considered: shapes, including small water bodies created in the

carbon dioxide (CO2): absent if values were lower parts of small water streams that drain the 1 o1.25 mg mL À corresponding to the detection abundant rainfall in the area. Most of the farms have 1 limit, present if values were 41.25 mg mL À ; running water almost all year round. A few have lim- 2 nitrites (NO À ): absent if values were ited water £ow or even depend exclusively on rain- 1 o0.05 mg mL À corresponding to the detection falls. The maximum depth is generally comprised 1 limit; present if values were 40.05 mg mL À ; between 1.5 and 4 m (Table 1) with variable £oating ammonia (NH41): absent if values were vegetation cover, and with shorelines generally cov- 1 o0.02 mg mL À corresponding to the detection ered by grass or other small plants, which were ex- limit; present if values were 40.02 mg mL; ceptionally steep. s pH and temperature were measured using a YSI The physicochemical parameters measured varied digital meter (Yellow Springs Instrument Com- within a small range (Table1)and were characteristic pany,Yellow Springs, OH, USA) at 15cm from the of good qualityand soft water. In two cases, pH values surface; attained relatively high levels (6.5 an 8.0), whereas 1 1 hardness (mg L À of CaCO3) levels o60 mg mL À the majority of other farms had pH values around 1 are considered to be low; 5.0 and hardness never exceeded 25 mg L À . These transparency was determined by the Secchi disc values correspond to acidic soft waters typical of method. these environments. Water transparency (Table 1) was generally low (mean of 33 cm). This low trans- parency is also typical of local waters and sometimes Hydromorphological parameters after rainfall the sediment content contributes to dra- The environmental parameters susceptible to play a matically decrease water transparency. Carbon diox- 2 41 role in the reproductive success of the paiche have ide, NO À and NH levels were, in most cases, been recorded in all farms. Six environmental para- below the detection limits of the measurement kit meters have been studied: pond area; pond maxi- and never reached critical values for ¢sh, indicating mum depth; presence of vegetation on pond water that rearing conditions were compatible with the surface; shape of the shoreline (slow or steep slope); known species preferenda. soil type (clay,sandy clay or sandy-muddy) and water (£ow through or stagnant water). Reproductive activity Field data collection and analysis The reproductive activity of A. gigas was determined We went to di¡erent ¢sh farms, and there we re- during the period 2007^2010 on the monthly num- quested to speak to the people who tend the paiches. ber of spawning events reported by the farmers. Then, we explained the reason for our presence, and Spawning generally occurred in a shallow area of requested permission to make measurements of their the pond in a small depression or nest dug by the pond water quality. During the interviews, we en- two brooders. Reproductions occurred all year quired about the total number of paiches and the round, although a marked seasonality was apparent food types and quantities distributed. During the vis- (Fig. 2). The maximum number of spawning events it, we also evaluated the total pond area, soil type and was observed between December and April (73 out water surface vegetation. of 100 spawnings over the 4-year period), corre- Data were analysed with the free statistical soft- sponding mainly to the rainy season, short photoper- iod and higher temperatures (Fig. 3). During the dry ware R, version 2.6.2 and Statgraphics Plus (StatPoint Technologies,Washington, DC, USA). season (May^September), reproductive activity (Fig. 2) was lower (27 out of100 spawnings).

Results Seasonality of fry production Pond characteristics The seasonality of fry production (Fig. 2) was more The visited farms are located along the road between pronounced than that of spawning activity. The Iquitos and Nauta (Fig. 1). Farms are generally monthly fry production peaked during October^

r 2011 Blackwell Publishing Ltd, Aquaculture Research, 42,815^822 817 Reproductive aspects of ARAPAIMA in captivity JNu¤ nì ez et al. Aquaculture Research, 2011, 42, 815^822

Figure 1 Aerial view of the‘Carretera Iquitos-Nauta’ Loreto, Peru. Points indicate the location of the main farms.

Table 1 Summary of some physicochemical and physical characteristics of Arapaima gigas breeding ponds in the region of Iquitos, Peru (Iquitos-Nauta road)

Carbon Hardness Transparency Temperature Area Depth 1 2 dioxide Nitrites Ammonia pH (mg mL À ) (cm) ( 1C) (m ) (m)

Minimum 2 0.05 0.2 4.25 7.0 7 27.5 500 1.5 Maximum 22 0.15 3 7.5 25.0 124 32.4 8 000 4 Mean 8.85 0.089 1.06 5.47 15.4 33 29.9 2 923 2.12 ( SD) 6.24 0.121 0.82 0.86 4.32 30 1.46 2 142 0.82 Æ

March, and a very limited fry production was April to November, but di¡ered strikingly from De- observed during the dry season from April to cember to March (Fig. 2). During this second half of September, corresponding to longer days and lower the maximum reproductive activity period, the temperatures (Fig. 3). Patterns of fry production and monthly number of ¢ngerlings per spawning de- number of fry per spawning were quite similar from creased remarkably to reach its minimum in April.

818 r 2011 Blackwell Publishing Ltd, Aquaculture Research, 42,815^822 Aquaculture Research, 2011, 42, 815^822 Reproductive aspects of ARAPAIMA in captivity JNu¤ nì ez et al.

45 40 35 30 25 20 15 10 5 0 Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May

Cumul. fry / month (x10–3) Monthly spawnings Monthly fry nr / spawning (x10–3)

Figure 2 Four-year (2007^2010) cumulated monthly number of Arapaima gigas spawnings (thick grey line), total fry (solid dark line) and fry number per spawning (dotted dark line) of16 farms along the Iquitos-Nauta road.Values must be multiplied by103 for total fry and fry per spawning.

35 760

30 C)

° 740

25 720

20

700

15

680 10 Photoperiod (min. of light)

660 5

Nr of spawnings0 Rainfall (cm) T ( 640 JUL JAN JUN FEB SEP APR OCT DEC NOV MAY AUG MAR

Figure 3 Monthly variations of rainfall (left axis, thick dark line), temperature (left axis, dotted dark line), cumulated (2007^2010) number of spawnings of Arapaima gigas females (left axis, light dark line) and photoperiod (right axis, dotted black line) in the region of Iquitos. Sources: IIAP (temperature and rainfall) and Bureau des Longitudes: http://www. bureau-des-longitudes.fr (photoperiod calculations for Iquitos).

In terms of ¢ngerling production per spawning, the Farm e⁄ciency on reproduction and fry best period corresponded to October^December; production nevertheless, the higher ¢ngerling production rate Total fry production, which was highly variable was observed in February^March (Fig. 2). among farms, reached over 13000 in certain farms

r 2011 Blackwell Publishing Ltd, Aquaculture Research, 42,815^822 819 Reproductive aspects of ARAPAIMA in captivity JNu¤ nì ez et al. Aquaculture Research, 2011, 42, 815^822

12000 1200

8000 800 Total fry 4000 400

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Fry / spawning Fry / female year Farms

Figure 4 Cumulated monthly fry number (left axis, open bars), fry per spawning (right axis, shaded bars) and fry per female per year (left axis, black bars) of Arapaima gigas over a 4-year period (2007^2010) for 16 farms along the road Iquitos-Nauta. during the 4-year period, while being o500 in fry number per spawning observed in this study.This others (Fig. 4) and was positively correlated with the situation can be explained by putative adverse ecolo- number of females (P 5 0.044) and the feeding level gical factors prevailing during the dry season in the (P 5 0.01). No signi¢cant correlation was found with ponds, which induced higher mortalities in larval other factors. The number of fry per spawning over and ¢ngerling stages. This means that even if the en- the 4-year period (Fig. 4) was highly variable too vironmental conditions are still favourable enough (208^1215), but was not correlated with feeding level for ¢sh reproduction, the conditions are not optimal (P 5 0.06). for larvae and ¢ngerling survival. During the rainy The total number of fry produced per female dur- season, maximum ¢ngerling survival is observed ing the 4-year period (results not shown) ranged during the ¢rst months of the £ood period (October^ from 106 to 4254, corresponding to 26^1063 ¢nger- December), but survival decreased progressively with lings per female and per year (Fig. 4) and this para- rainfall decline. As no speci¢c study has been com- meter was correlated with feeding level (P 5 0.02), pleted on plankton or other available food sources but not with other physicochemical or hydromor- for larvae feeding, it is di⁄cult to conclude on the pre- phological factors. dominance of the feeding factor over other environ- mental variables on larvae survival. The lack of information on the real fecundity of fe- males also hampers the interpretation of the data, as Discussion the average number of spawned mayalso be low- During this study, spawning generally occurred in a er during the dry season. In the Iquitos-Nauta area, shallow area of the pond in a small depression or nest the maximum number of fry collected 1 month after dug by the two brooders, which corresponded to pre- hatching from a single spawning was 3941, which is vious description in the wild (Bard & Imbiriba, 1986; considerably less than the maximum fecundity eval- Castello,2008). One of the most important ¢ndings of uated by various authors (reviewed by Imbiriba, 94). this study is the almost continuous reproductive ac- The reported absolute fecundities ranged from tivity of A. gigas reared in pond conditions, while this 58000 to 86000 maturing oocytes in wild ¢sh or a species is reported as an annual spawner with the maximum of 11000 juveniles in pond conditions in capability of few spawning events during the rainy Brazil (Bard & Imbiriba, 1986). From previous obser- season in Peru (Guerra, 1980) and Brazil (Imbiriba, vations (Godinho et al. 2005; Nu¤ nì ez & Duponchelle, 1991). Nevertheless, even if these results demonstrate 2009), the oocyte development of A. gigas is group that a residual reproductive activity persists during synchronous with three vitellogenic oocyte cohorts the dry season, the overall survival of ¢ngerlings on at di¡erent stages of development in the ovaries of harvest, 1 month in average after spawning, is very maturing females. This means that the e¡ective fe- weak during the dry season as attested by the low cundity per spawning is probably closer to one-third

820 r 2011 Blackwell Publishing Ltd, Aquaculture Research, 42,815^822 Aquaculture Research, 2011, 42, 815^822 Reproductive aspects of ARAPAIMA in captivity JNu¤ nì ez et al. of the reported maximum fecundity or could be properly fertilizing the ponds to allow optimal pri- equivalent to the oocyte number of the larger oocyte mary and secondary production. diameter cohort. Even if one takes into account this The results described in this study emphasize the hypothesis (i.e. around 20 000 spawned eggs), the to- high impact of feeding level on reproductive activity tal number of hatched larvae is very far below this and probablyalso on the survival of young stages and ¢gure in view of our observations during this 4-year it appeared that in most farms the feeding level of the study. The probability that such high rates of eggs breeders was insu⁄cient. The importance of nutri- and young stages mortality exist in a species provid- tional aspects has been also reported by (Saavedra ing elaborated bi-parental care seems unlikely. This Rojas, Quintero Pinto, Lopez Hernandez et al. 2005) species may experience high rates of atresia just be- for the quantity and quality of fry produced in captiv- fore spawning or poor survival during the embry- ity,but this is the ¢rst experimental observation of di- ogenesis stage, occurring in the nest at the bottom of rect in£uence of the feeding level on reproductive the pond, which might partly account for why esti- activity of A. gigas. In the Iquitos-Nauta area, feeding mated fecundities on ovaries in ¢nal maturation of A. gigas is achieved mainly with frozen or live ¢sh, stage are much higher than the observed number of this feed appeared to be e¡ective for breeding stocks guarded fry. Clearly, additional work is needed to at 2% of biomass per day, although a more detailed shed light on this issue and to better understand the study is needed to establish the exact amount and reproductive potential of theA. gigas in captivity. quality of food that would allow for optimal repro- The number of fry produced per female and per year ductive rate and fry survival in this species. has been positively correlated with the feeding level; In conclusion, this study shows that under good the farms where the food distribution was very low ex- management conditions including su⁄cient food hibited also a lower number of spawning. Among all supply, the production of A. gigas ¢ngerlings in cap- other factors studied, none appeared to play a key role tivity can be substantial (more than 1000 fry per fe- on the number of spawning or the total number of fry male per year) and holds a good potential for produced. In particular, the presence or absence of improvement, despite the low fecundity of this spe- £oating vegetation seemed to have no in£uence in the cies. These results also substantiate the possibility of reproductive activity of A. gigas in captivity. Concern- sustained fry production almost during 8 months ing NH41, the absence of correlation with the number year round from September to April. Finally, our of spawning indicated that NH41 had no direct e¡ect study con¢rms the possibility of aquaculture devel- within the measured range on female reproductive ac- opment of A. gigas through an improvement of tivity,and did not a¡ect ¢sh behaviour becauseA. gigas broodstock management. can tolerate much higher concentrations than those observed in this study (Cavero, Pereira-¢lho, Bordin- hon, Fonseca, Ituassu¤ , Roubach & Ono 2004). Acknowledgements Regardless of the number of spawning events, the This study was carried out as part of the scienti¢c quantity of fry produced per brood varied greatly collaborations between the IIAP and the Institut de from one farm to another. These di¡erences can be Recherche pour le De¤ veloppement (IRD), both part explained by lower fecundities or di¡erential mortal- of the research network known as RIIA (Red de In- ities during the1-month average period before ¢nger- vestigacio¤ n sobre la Ictiofauna Amazo¤ nica, http:// ling harvest. The most likely hypothesis to explain www.riiaamazonia.org). these variable mortalities is that feeding of fry de- The authors wish to thank all the farm owners and pends greatly on the productivity of ponds. As juve- people who contributed to the ¢eld work. niles and ¢ngerlings feeding in the wild consists This work has been funded in part by the Peruvian mainly of zooplankton, insects, insect larvae and research programme INCAGRO. small gastropods (Oliveira, Poleto & Venere 2005), This is a publication IRD-DIVA-ISEM 2011-042. the main causes of mortality are probably an insu⁄- cient food availability. This hypothesis needs to be tested by a detailed study of the limnological para- meters and plankton productivity in the di¡erent Reference farming environments. If con¢rmed, considering the Alcantara F. & Guerra H. (1992) Cultivo de paiche, Arapaima huge variations observed on harvested fry among gigas, utilizando bujurqui, Cichlasoma bimaculatum, como farms, ¢ngerling production could be improved by presa. Folia Amazo¤ nica 4,129^139.

r 2011 Blackwell Publishing Ltd, Aquaculture Research, 42,815^822 821 Reproductive aspects of ARAPAIMA in captivity JNu¤ nì ez et al. Aquaculture Research, 2011, 42, 815^822

Bard J. & Imbiriba E.P.(1986) Piscicultura de pirarucu, Arapai- Imbiriba E.P.(1991) Produc° aì ao e manjo de alevinos de pirarucu, ma gigas. No.52. EMBRAPA-CPATU, Bele¤ m, Brazil pp.52. Arapaima gigas (Cuvier).Technical Report EMBRAPA-CPA- Castello L. (2008) Nesting habitat of Arapaima gigas (Schinz) TU,57,19pp. in Amazonian £oodplains. Journal of Fish Biology 72, Nunì ez J. (2009) Domestication de nouvelles espe' ces d’inte¤ reŒ t 1520^1528. aquacole en Amazonie. Cahiers Agriculture18,136^143. Cavero B.A.S., Pereira-Filho M., Bordinhon A.M., Fonseca Nu¤ nì ez J. & Duponchelle F. (2009) Towards a universal scale F.A.L., Ituassu¤ D.R., Roubach R. & Ono E.A. (2004) Toler- to assess sexual maturation and related life history traits a“ ncia de juvenis de pirarucu ao aumento da concentrac° aì o in oviparous ¢shes. Fish Physiology and Biochemis- de amoŒ nia em ambiente con¢nado. Pesquisa Agropecua?r- try 35,167^180. ia Brasileira 39, 513–516. OliveiraV.D., Poleto S.L. & Venere P.C. (2005) Feeding of juve- Chu-Koo F. & AlcaŁ ntara F. (2009) Paiche dome¤ stico en la nile pirarucu (Arapaima gigas, ) in their nat- Amazon|¤a: perspectivas de una crianza sostenible. Pesca ural environment, lago Quatro Bocas, Araguaiana-MT, Responsable 57,32^33. Brazil. Neotropical Ichthyology 3,312^314. Chu-Koo F.& Tello S. (2010) Produccio¤ n de semilla de Paiche Pereira Filho M. & Roubach R. (2005) Pirarucu, Arapaima en Peru¤ . Infopesca Internacional 41,30^35. gigas. In Espe¤ cies nativas para piscicultura no Brasil (ed. Chu-Koo F., Dugue¤ R., AlvaŁ n Aguilar M., Casanova Daza A., by H. Rosa Nascimento), pp. 37–62. UFSM, Santa AlcaŁ ntara Bocanegra F., ChaŁ vezVeintemilla C., Duponch- Maria, Brasil. elle F., Renno J.F., Tello S. & Nunì ez J. (2009) Gender Rebaza M., Rebaza C. & Deza S. (2010) Densidad de siembra determination in the Paiche or Pirarucu (Arapaima gigas) para cultivos de Paiche en jaulas £otantes. Aquavisio¤ n 6, using plasma vitellogenin, 17b-estradiol, and 11-ketotes- 26^27. tosterone levels. Fish Physiology and Biochemistry 35, Saavedra Rojas E.A., Quintero Pinto L.G. & Landines Parra M.A. 125^136. (2005) Aspectos reproductivos. In: Biolog|¤ a y cultivo del pir- Godinho H.P., Santos J.E., Formagio P.S. & Guimaraes-Cruz arucu¤ Arapaima gigas Schinz, 1822. Pisces: Arapaimidae. R.J. (2005) Gonadal morphology and reproductive traits Bases para un aprovechamiento sostenible ^ Aspectos Repro- of the Amazonian ¢sh Arapaima gigas (Schinz,1822). Acta ductivos (ed. by A.I. Sanabria, I.C. Beltran & P.V. Zoologica 86,289^294. Daza), pp. 31–40. INCODER/UNC, Bogota´ , Colombia. Guerra F.H. (1980) Desarollo sexual del paiche, Arapaima Saavedra Rojas E.A., Quintero Pinto L.G., Lopez Hernandez gigas, en las zonas reservadas del estado (R|¤o Pacaya Sa- N. & Pezzato L.E. (2005) Nutricio¤ n y alimentacio¤ n del pir- miria) 1971^1975. In : Informe IMARPE No. 67, pp. 20. arucu¤ Arapaima gigas (Schinz,1882). In Biolog|¤ a y cultivo Instituto del Mar del Peru´ , Callao, Peru´ . del pirarucu¤ Arapaima gigas Schinz,1822. Pisces: Arapaimi- Guerra H., AlcaŁ ntara F., Padilla P., Rebaza M.,Tello S., Ismi- dae. Bases para un aprovechamiento sostenible ^ Aspectos nì o R., Rebaza C., Deza S., Ascon G., Iberico J., MontreuilV. Reproductivos (ed. by A.I. Sanabria, I.C. Beltran & & Limachi L. (2002) Produccio¤ n y manejo de alevinos de P.V. Daza), pp. 41–58. INCODER/UNC, Bogota´ , paiche. : IIAP,Iquitos, Peru¤ . Colombia.

822 r 2011 Blackwell Publishing Ltd, Aquaculture Research, 42,815^822