**6. Larval diversity**

Eggs and larvae identification depends on the integrity of the sampled larvae as well as on previous ontogeny studies. High diversity of the Amazon fish and the paucity of ichthyo‐ plankton ontogeny studies become the larvae identification a challenge. Despite of this, just 0.3% of the larvae and juveniles collected were completely unknown and only 8% of them were identified at the order level, the lowest possible level. The most larvae (92%) was identified at family level and the identification at genus and species level was possible in 30% and 13% of the total larvae, respectively. The Characiformes and Siluriformes larvae were dominant in the samples and represented about 98% of the total larvae collected. The PL net method collected almost twofold more Characiformes than Siluriformes larvae, while IJ method collected almost fourfold Siluriformes than Characiformes larvae. The larvae composition collected by IL method did not show a large discrepancy among the most important taxonomic groups, even so Characiformes were more abundant than Siluriformes larvae (Table 6).

**Order Total IJ IL PL**

http://dx.doi.org/10.5772/57404

149

Diversity and Abundance of Fish Larvae Drifting in the Madeira River, Amazon Basin: Sampling Methods Comparison

Unknown 63 0,3% 0 0% 29 0.4% 34 0.3% Characiformes 12,485 58% 231 2% 4,005 55% 8,249 62% Siluriformes 8,806 41% 825 74% 3,182 44% 4,799 36% Perciformes 189 1% 20 2% 49 1% 120 1% Clupeiformes 69 0% 12 1% 21 0.3% 36 0.3% Gymnotiformes 53 0% 32 3% 9 0.1% 12 0.1%

Unknown 1,834 8% 27 2% 832 11% 975 7% Curimatidae 5,891 27% 9 1% 1,676 23% 4,206 32% Auchenipteridae 3,965 18% 11 1% 1,272 17% 2,682 20% Pimelodidae 3,960 18% 717 64% 1,582 22% 1,661 13% Characidae 2,650 12% 134 12% 880 12% 1,636 12% Prochilodontidae 1,482 7% 1 0,1% 495 7% 986 7% Anostomidae 749 3% 1 0,1% 231 3% 517 4% Hemiodontidae 276 1% <0.1% 101 1% 175 1% Cynodontidae 251 1% 71 6% 70 1% 110 1% Sciaenidae 188 1% 20 2% 48 1% 120 1% Doradidae 116 1% 67 6% 20 0.3% 29 0.2% Trichomycteridae 92 0.4% 5 0.4% 34 0.5% 53 0.4% Cetopsidae 61 0.3% 3 0.3% 16 0.2% 42 0.3% Heptapteridae 45 0.2% 3 0.3% 16 0.2% 26 0.2% Apteronotidae 27 0.1% 22 2% 2 <0.1 3 <0.1 Callichthyidae 15 0.1% 34 1% 2 <0.1 5 <0.1 Loricariidae 15 0.1% 16 0.1% 7 0.1% 7 0.1% Engraulidae 13 0.1% 9 1% 3 <0.1 1 <0.1 Pristigasteridae 11 0.1% 3 0.3% 2 <0.1 6 <0.1 Erythrinidae 8 <0.1 2 0.3% 4 0.1% 2 <0.1 Gasteropelecidae 8 <0.1 0 <0.1 2 <0.1 6 <0.1 Sternopygidae 7 <0.1 6 1% 0 0% 1 <0.1 Clupeidae 1 <0.1 0 0% 0 0% 1 <0.1 Total 21,665 1 1,120 100% 7,295 100% 13,250 100%

**Table 6.** Larvae composition caught by the combination of methods and nets: IJ- integrating sampling with juvenile

net; IL- integrating sampling with larval net; and PL- point sampling with larval net.

**Order**

**Family**

Curimatidae,AuchenipteridaeandPimelodidaewerethemostabundantfamiliesthattotalizing about 2/3 of the total larvae. The Curimatidae and Auchenipteridae families represented the half of the larvae collected by the PL method and the Pimelodidae family represented 2/3 of the collected by IJ method. The caught of three families aforementioned were similar for the IL method. The Characidae family was the fourth in the rank of abundance, representing about 12% of the larval composition collected by the three methods. The larvae of Prochilodontidae, Anostomidae and Hemiodontidae together represented 12% of the total larvae. The sampling methods using larval net collected 95% or more of the larvae of Characidae, Prochilodonti‐ dae, Anostomidae and Hemiodontidae families, which they represented together about ¾ of all larvae caught. Other families represented less than 5% of the total (Table 6).

Less than 5% of the larvae of Curimatidae, Auchenipteridae, Anostomidae, Prochilodontidae and Hemiodontidae families, the most abundant families, were identified at the genus level. Nonetheless, 94% of the Characidae larvae and 83% of the Pimelodidae were identified at the genus level. The larvae of *Brycon*, *Mylossoma*, *Triportheus* and *Piaractus* genera were the most abundant among the Characidae family. Each one represented more than 5% of all Characidae larvae and together represented 93% of the Characidae larvae and 11% of all larvae, but few of them were identified at the species level. On the other hand, most larvae of the four most abundant genera of the Pimelodidae family were identified at the species level. *Pinirampus*, *Brachyplatystoma*, *Pimelodus* and *Pseudoplatystoma* represented each one more than 5% of all Pimelodidae larvae and together represented 71% of Pimelodidae larvae and 13% of all larvae.

The Curimatidae, Auchenipteridae and Anostomidae families are diversified and they were represented in the study area by 7 genera and 21 species, 18 genera and 21 species and 6 genera and 13 species, respectively. For the other side, the diversity of Prochilodontidae and Hemio‐ dontidae families are low and they have only two genera with three species and two genera with six species in the area, respectively [22].

[22] identified three species of *Brycon* (*B. amazonicus, B. falcatus* and *B. melanopterus*), two species of *Mylossoma* (*M. aureum* and *M. duriventre*) and three species of *Triportheus* (*T. albus, T.* Diversity and Abundance of Fish Larvae Drifting in the Madeira River, Amazon Basin: Sampling Methods Comparison http://dx.doi.org/10.5772/57404 149

**6. Larval diversity**

148 Biodiversity - The Dynamic Balance of the Planet

Eggs and larvae identification depends on the integrity of the sampled larvae as well as on previous ontogeny studies. High diversity of the Amazon fish and the paucity of ichthyo‐ plankton ontogeny studies become the larvae identification a challenge. Despite of this, just 0.3% of the larvae and juveniles collected were completely unknown and only 8% of them were identified at the order level, the lowest possible level. The most larvae (92%) was identified at family level and the identification at genus and species level was possible in 30% and 13% of the total larvae, respectively. The Characiformes and Siluriformes larvae were dominant in the samples and represented about 98% of the total larvae collected. The PL net method collected almost twofold more Characiformes than Siluriformes larvae, while IJ method collected almost fourfold Siluriformes than Characiformes larvae. The larvae composition collected by IL method did not show a large discrepancy among the most important taxonomic groups, even

Curimatidae,AuchenipteridaeandPimelodidaewerethemostabundantfamiliesthattotalizing about 2/3 of the total larvae. The Curimatidae and Auchenipteridae families represented the half of the larvae collected by the PL method and the Pimelodidae family represented 2/3 of the collected by IJ method. The caught of three families aforementioned were similar for the IL method. The Characidae family was the fourth in the rank of abundance, representing about 12% of the larval composition collected by the three methods. The larvae of Prochilodontidae, Anostomidae and Hemiodontidae together represented 12% of the total larvae. The sampling methods using larval net collected 95% or more of the larvae of Characidae, Prochilodonti‐ dae, Anostomidae and Hemiodontidae families, which they represented together about ¾ of

Less than 5% of the larvae of Curimatidae, Auchenipteridae, Anostomidae, Prochilodontidae and Hemiodontidae families, the most abundant families, were identified at the genus level. Nonetheless, 94% of the Characidae larvae and 83% of the Pimelodidae were identified at the genus level. The larvae of *Brycon*, *Mylossoma*, *Triportheus* and *Piaractus* genera were the most abundant among the Characidae family. Each one represented more than 5% of all Characidae larvae and together represented 93% of the Characidae larvae and 11% of all larvae, but few of them were identified at the species level. On the other hand, most larvae of the four most abundant genera of the Pimelodidae family were identified at the species level. *Pinirampus*, *Brachyplatystoma*, *Pimelodus* and *Pseudoplatystoma* represented each one more than 5% of all Pimelodidae larvae and together represented 71% of Pimelodidae larvae and 13% of all larvae.

The Curimatidae, Auchenipteridae and Anostomidae families are diversified and they were represented in the study area by 7 genera and 21 species, 18 genera and 21 species and 6 genera and 13 species, respectively. For the other side, the diversity of Prochilodontidae and Hemio‐ dontidae families are low and they have only two genera with three species and two genera

[22] identified three species of *Brycon* (*B. amazonicus, B. falcatus* and *B. melanopterus*), two species of *Mylossoma* (*M. aureum* and *M. duriventre*) and three species of *Triportheus* (*T. albus, T.*

so Characiformes were more abundant than Siluriformes larvae (Table 6).

all larvae caught. Other families represented less than 5% of the total (Table 6).

with six species in the area, respectively [22].


**Table 6.** Larvae composition caught by the combination of methods and nets: IJ- integrating sampling with juvenile net; IL- integrating sampling with larval net; and PL- point sampling with larval net.

*angulatus* and *T. auritus*) in the study area. The present study identified the two species of *Mylossoma* and only two species of Triportheus (*T. angulatus* and *T. auritus*) and none of *Brycon*. Considering the shortage of larvae identified at specie level, it is not possible to discuss the composition of those genera. On the other side, all *Pinirampus* larvae and 99% of the *Brachyplatystoma* larvae were identified at specie level. *Pinirampus pirinampu* was the only specie of the genus. Five *Brachyplatystoma* species were identified, being *B. filamentosum, B. rousseauxii* and *B. capapretum* the most abundant, totalizing 90% of the genus. The other two species in the larval collection were *B. platynemum* and *B. juruense*. Two *Brachyplatystoma* species, *B. tigrinum* and *B. vaillantii*, were present in the local fish collection [22] but they are not present in the present larvae samples. Nevertheless, only 56% of the *Pimelodus* and 11% of the *Pseudoplatystoma* genus were identified at specie level. *Pimelodus blochii* was the main specie identified of the genus and only one individual represented *P. altissimus* was in all samples. The two species of *Pseudoplatystoma* present in the area were in the larvae samples, *P. puncti‐ fer* and *P. tigrinum*, but most larvae were not identified at specie level. The larval net caught the majority of the larvae of those species, with the exceptions of *P. blochii, B. rousseauxii* and *B. capapretum*, which the juvenile net caught more than 50% of the larvae of each specie.

discharge months (December-January), and *Mylossoma* spp. extended the reproduction beyond the period of the rising discharge (December-March). *Triportheus* spp. presented a delayed drifting larval movement in comparison with the other species. The peak was detected in March by the IL method and in May by the PL method, both moments in the high discharge

Diversity and Abundance of Fish Larvae Drifting in the Madeira River, Amazon Basin: Sampling Methods Comparison

0

100

200

**Larval Flux** 300

400

1 2 3 4 5 6 7 8 9 10 11 12 **Month**

1 2 3 4 5 6 7 8 9 10 11 12 **Month**

IJ IL PL

**Anostomidae**

1 2 3 4 5 6 7 8 9 10 11 12 **Mont hs**

IJ IL PL

**Characidae**

**Auchenipteridae**

http://dx.doi.org/10.5772/57404

151

IJ IL PL

Figure 6. Monthly average of the larval flux (larvae/s) of main families: IJ- integrating sampling with juvenile net; ILintegrating sampling with larval net; and PL- point sampling with larval net. Months: 2-5 (February-May) high discharge; 6-7 (June-July) falling discharge; 8-11 (August-November) low discharge; 12-1 (December-January) rising

**Figure 6.** Monthly average of the larval flux (larvae/s) of main families: IJ- integrating sampling with juvenile net; ILintegrating sampling with larval net; and PL- point sampling with larval net. Months: 2-5 (February-May) high dis‐ charge; 6-7 (June-July) falling discharge; 8-11 (August-November) low discharge; 12-1 (December-January) rising

0

50

100

**Larval Flux** 150

0

100

**Larval Flux**

200

300

period (Figure 7).

**Larval Flux**

> 1 2 3 4 5 6 7 8 9 10 11 12 **Month**

> 1 2 3 4 5 6 7 8 9 10 11 12 **Month**

**Prochilodontidae**

1 2 3 4 5 6 7 8 9 10 11 12 **Month**

IJ IL PL

**Pimelodidae**

IJ IL PL

**Curimatidae**

IJ IL PL

discharge.

discharge.

0

50

100

**Larval Flux**

150

200

0

100

**Larval Flux**

200

300

The majority (c.a. 95%) of the larvae of the most families were in pre-flexion or earlier stages, whilst Characidae and Pimelodidae presented the larvae in advanced stage. The most Characidae larvae were found in pre-flexion (63%) and flexion (35%) stages, while Pimelodidae larvae were found in flexion (69%) and post-flexion (23%) stages. However, the development stage composition was different for each species or genus. *Mylossoma* spp., *Triportheus* spp. and *Piaractus brachypomus* presented 90% or more of the larvae in pre-flexion stage, while *Brycon* spp. showed developed larvae, with 87% in flexion stage. The majority (2/3 or more) of *Pseudoplatystoma* spp., *Zungaro zungaro, Pinirampus pirinampu*, *Brachyplatystoma filamentosum* and *B. capapretum* larvae were in flexion stage, and more than a half of *Pimelodus blochii* and *B. rousseauxii* were in post-flexion stage.

The larvae drift in the river occurred mainly when the discharge is increasing. Some families concentrated the drifting movement in just a few months and others remained drifting during all year, indicating the length of the reproductive season. The average larval flux estimated by the three samples method in the two sites was compared in order to understand the general drifting pattern of the most abundant taxa. The IL and PL methods were efficient to identify the great variation of the larval flux for most of the families and IJ detected satisfactorily the annual flux variation only for Characidae and Pimelodidae.

The Curimatidae and Prochilodontidae families showed the shortest reproductive season between December and April, but strongly concentrated in March. The larvae of Auchenip‐ teridae and Anostomidae families remained drifting in the river for a longer period, between November and May, with peak in January and March. The general pattern of the larval drifting of the Characidae and Pimelodidae families is also intensive in the period of rising discharge and less intensive in the low discharge months. However, the results of the three methods showed some differences, especially in April and May, when the IL method pointed the end of reproduction season and the PL method still detects an intensive flux of larvae (Figure 6). *Brycon* spp. and *Piaractus brachypomus* showed a shorter spawning season, during the rising discharge months (December-January), and *Mylossoma* spp. extended the reproduction beyond the period of the rising discharge (December-March). *Triportheus* spp. presented a delayed drifting larval movement in comparison with the other species. The peak was detected in March by the IL method and in May by the PL method, both moments in the high discharge period (Figure 7).

*angulatus* and *T. auritus*) in the study area. The present study identified the two species of *Mylossoma* and only two species of Triportheus (*T. angulatus* and *T. auritus*) and none of *Brycon*. Considering the shortage of larvae identified at specie level, it is not possible to discuss the composition of those genera. On the other side, all *Pinirampus* larvae and 99% of the *Brachyplatystoma* larvae were identified at specie level. *Pinirampus pirinampu* was the only specie of the genus. Five *Brachyplatystoma* species were identified, being *B. filamentosum, B. rousseauxii* and *B. capapretum* the most abundant, totalizing 90% of the genus. The other two species in the larval collection were *B. platynemum* and *B. juruense*. Two *Brachyplatystoma* species, *B. tigrinum* and *B. vaillantii*, were present in the local fish collection [22] but they are not present in the present larvae samples. Nevertheless, only 56% of the *Pimelodus* and 11% of the *Pseudoplatystoma* genus were identified at specie level. *Pimelodus blochii* was the main specie identified of the genus and only one individual represented *P. altissimus* was in all samples. The two species of *Pseudoplatystoma* present in the area were in the larvae samples, *P. puncti‐ fer* and *P. tigrinum*, but most larvae were not identified at specie level. The larval net caught the majority of the larvae of those species, with the exceptions of *P. blochii, B. rousseauxii* and *B. capapretum*, which the juvenile net caught more than 50% of the larvae of each specie.

The majority (c.a. 95%) of the larvae of the most families were in pre-flexion or earlier stages, whilst Characidae and Pimelodidae presented the larvae in advanced stage. The most Characidae larvae were found in pre-flexion (63%) and flexion (35%) stages, while Pimelodidae larvae were found in flexion (69%) and post-flexion (23%) stages. However, the development stage composition was different for each species or genus. *Mylossoma* spp., *Triportheus* spp. and *Piaractus brachypomus* presented 90% or more of the larvae in pre-flexion stage, while *Brycon* spp. showed developed larvae, with 87% in flexion stage. The majority (2/3 or more) of *Pseudoplatystoma* spp., *Zungaro zungaro, Pinirampus pirinampu*, *Brachyplatystoma filamentosum* and *B. capapretum* larvae were in flexion stage, and more than a half of *Pimelodus blochii* and

The larvae drift in the river occurred mainly when the discharge is increasing. Some families concentrated the drifting movement in just a few months and others remained drifting during all year, indicating the length of the reproductive season. The average larval flux estimated by the three samples method in the two sites was compared in order to understand the general drifting pattern of the most abundant taxa. The IL and PL methods were efficient to identify the great variation of the larval flux for most of the families and IJ detected satisfactorily the

The Curimatidae and Prochilodontidae families showed the shortest reproductive season between December and April, but strongly concentrated in March. The larvae of Auchenip‐ teridae and Anostomidae families remained drifting in the river for a longer period, between November and May, with peak in January and March. The general pattern of the larval drifting of the Characidae and Pimelodidae families is also intensive in the period of rising discharge and less intensive in the low discharge months. However, the results of the three methods showed some differences, especially in April and May, when the IL method pointed the end of reproduction season and the PL method still detects an intensive flux of larvae (Figure 6). *Brycon* spp. and *Piaractus brachypomus* showed a shorter spawning season, during the rising

*B. rousseauxii* were in post-flexion stage.

150 Biodiversity - The Dynamic Balance of the Planet

annual flux variation only for Characidae and Pimelodidae.

Figure 6. Monthly average of the larval flux (larvae/s) of main families: IJ- integrating sampling with juvenile net; ILintegrating sampling with larval net; and PL- point sampling with larval net. Months: 2-5 (February-May) high discharge; 6-7 (June-July) falling discharge; 8-11 (August-November) low discharge; 12-1 (December-January) rising discharge. **Figure 6.** Monthly average of the larval flux (larvae/s) of main families: IJ- integrating sampling with juvenile net; ILintegrating sampling with larval net; and PL- point sampling with larval net. Months: 2-5 (February-May) high dis‐ charge; 6-7 (June-July) falling discharge; 8-11 (August-November) low discharge; 12-1 (December-January) rising discharge.

Figure 7. Monthly average of the larval flux (larvae/s) of the main species of Characidae family: IJ- integrating sampling with juvenile net; IL- integrating sampling with larval net; and PL- point sampling with larval net. Months: 2- 5 (February-May) high discharge; 6-7 (June-July) falling discharge; 8-11 (August-November) low discharge; 12-1 (December-January) rising discharge. **Figure 7.** Monthly average of the larval flux (larvae/s) of the main species of Characidae family: IJ- integrating sam‐ pling with juvenile net; IL- integrating sampling with larval net; and PL- point sampling with larval net. Months: 2-5 (February-May) high discharge; 6-7 (June-July) falling discharge; 8-11 (August-November) low discharge; 12-1 (De‐ cember-January) rising discharge.

The drifting movement of the most abundant Pimelodidae species was longer than the Characidae species. In addition, the IJ method presented an expressive flux estimative when compared with the other groups, in special for the *Brachyplatystoma rousseauxii* and *Pimelodus blochii*. Despite of large amount of larvae of each species, drifting in the channel during the rising and or high discharge months, the movement of larvae did not stop during the low discharge period suggesting an uninterrupted spawning season. Larvae of *Pseudoplatystom*a spp. and *P. blochii* were scarce during the falling and low discharge, but the larvae of *Pinirampus pirinampu* or *Brachyplatystoma* spp. were abundant in that period. The different sampling methods also presented expressive differences in the larval flux of some Pimelodidae species. *P. blochii, B. capapretum* and *B. rousseauxii* presented isolated peaks of larval flux detected by only one method (Figure 6). The drifting movement of the most abundant Pimelodidae species was longer than the Characidae species. In addition, the IJ method presented an expressive flux estimative when compared with the other groups, in special for the *Brachyplatystoma rousseauxii* and *Pimelodus blochii*. Despite of large amount of larvae of each species, drifting in the channel during the rising and or high discharge months, the movement of larvae did not stop during the low discharge period suggesting an uninterrupted spawning season. Larvae of *Pseudoplatystom*a spp. and *P. blochii* were scarce during the falling and low discharge, but the larvae of *Pinirampus pirinampu* or *Brachyplatystoma* spp. were abundant in that period. The different sampling methods also presented expressive differences in the larval flux of some Pimelodidae species. *P. blochii, B. capapretum* and *B. rousseauxii* presented isolated peaks of larval flux detected by only one method (Figure 6).

#### **7. Discussion**

The point sampling method carried out at fixed habitats along of transects is a good way to compare the larval abundance or composition in different depths or habitats [2, 3, 4, 5, 23]. However, when the aim is to assess the impact of the river modification over the ichthyo‐ plankton drifting process, it is necessary to develop a methodology adapted to estimate the

The point sampling method carried out at fixed habitats along of transects is a good way to compare the larval abundance or composition in different depths or habitats [2, 3, 4, 5, 23]. However, when the aim is to assess the impact of the river modification over the ichthyoplankton drifting process, it is necessary to develop a methodology adapted to estimate the larval flux in natural or dammed river cross sections.

Figure 8. Monthly average of the larval flux (larvae/s) of the main species of Pimelodidae family: IJ- integrating sampling with juvenile net; IL- integrating sampling with larval net; and PL- point sampling with larval net. Months: 2- 5 (February-May) high discharge; 6-7 (June-July) falling discharge; 8-11 (August-November) low discharge; 12-1

**Figure 8.** Monthly average of the larval flux (larvae/s) of the main species of Pimelodidae family: IJ- integrating sam‐ pling with juvenile net; IL- integrating sampling with larval net; and PL- point sampling with larval net. Months: 2-5 (February-May) high discharge; 6-7 (June-July) falling discharge; 8-11 (August-November) low discharge; 12-1 (De‐

0

20

**Larval Flux** 40

1 2 3 4 5 6 7 8 9 10 11 12 **Month**

1 2 3 4 5 6 7 8 9 10 11 12 **Month**

> IJ IL PL

*B. capapretum*

1 2 3 4 5 6 7 8 9 10 11 12 **Mont h**

IJ IL PL

http://dx.doi.org/10.5772/57404

153

IJ IL PL

*P. blochii*

*Pseudoplatystoma* **spp.**

0

0

50

100

**Larval Flux** 150

50

**Larval Flux**

Diversity and Abundance of Fish Larvae Drifting in the Madeira River, Amazon Basin: Sampling Methods Comparison

1 2 3 4 5 6 7 8 9 10 11 12 **Month**

*B. filamentosum*

IJ IL PL

1 2 3 4 5 6 7 8 9 10 11 12 **Month**

*B. rousseauxii*

IJ IL PL

IJ IL PL

*P. pirinampu*

0

0

50

**Larval Flux** 100

50

100

**Larval Flux** 150

200

100

The three methods caught few eggs in relation to total of ichthyoplankton, and it was similar for the up and downriver sites. [4] also found few eggs in the ichthyoplankton composition in the Solimões River (Table 2). They hypothesized that the eggs have a short residence time, less than 16 hours, and as the spawning occurs mainly during the dusk, the eggs had hatched before

The three methods caught few eggs in relation to total of ichthyoplankton, and it was similar for the up and downriver sites. [4] also found few eggs in the ichthyoplankton composition in the Solimões River

larval flux in natural or dammed river cross sections.

1 2 3 4 5 6 7 8 9 10 11 12 **Mont h**

(December-January) rising discharge.

**7. Discussion** 

cember-January) rising discharge.

0

20

**Larval Flux** 40

Diversity and Abundance of Fish Larvae Drifting in the Madeira River, Amazon Basin: Sampling Methods Comparison http://dx.doi.org/10.5772/57404 153

Figure 7. Monthly average of the larval flux (larvae/s) of the main species of Characidae family: IJ- integrating sampling with juvenile net; IL- integrating sampling with larval net; and PL- point sampling with larval net. Months: 2- 5 (February-May) high discharge; 6-7 (June-July) falling discharge; 8-11 (August-November) low discharge; 12-1

**Figure 7.** Monthly average of the larval flux (larvae/s) of the main species of Characidae family: IJ- integrating sam‐ pling with juvenile net; IL- integrating sampling with larval net; and PL- point sampling with larval net. Months: 2-5 (February-May) high discharge; 6-7 (June-July) falling discharge; 8-11 (August-November) low discharge; 12-1 (De‐

0

20

40

**Larval Flux** 60

80

0

50

**Larval Flux** 100

1 2 3 4 5 6 7 8 9 10 11 12 **Month**

*P. brachypomus*

1 2 3 4 5 6 7 8 9 10 11 12 **Mont h**

 IJ IL PL

*Mylossoma* **spp.**

IJ IL PL

The drifting movement of the most abundant Pimelodidae species was longer than the Characidae species. In addition, the IJ method presented an expressive flux estimative when compared with the other groups, in special for the *Brachyplatystoma rousseauxii* and *Pimelodus blochii*. Despite of large amount of larvae of each species, drifting in the channel during the rising and or high discharge months, the movement of larvae did not stop during the low discharge period suggesting an uninterrupted spawning season. Larvae of *Pseudoplatystom*a spp. and *P. blochii* were scarce during the falling and low discharge, but the larvae of *Pinirampus pirinampu* or *Brachyplatystoma* spp. were abundant in that period. The different sampling methods also presented expressive differences in the larval flux of some Pimelodidae species. *P. blochii, B. capapretum* and *B. rousseauxii* presented isolated peaks of larval flux detected by only one method (Figure

The drifting movement of the most abundant Pimelodidae species was longer than the Characidae species. In addition, the IJ method presented an expressive flux estimative when compared with the other groups, in special for the *Brachyplatystoma rousseauxii* and *Pimelodus blochii*. Despite of large amount of larvae of each species, drifting in the channel during the rising and or high discharge months, the movement of larvae did not stop during the low discharge period suggesting an uninterrupted spawning season. Larvae of *Pseudoplatystom*a spp. and *P. blochii* were scarce during the falling and low discharge, but the larvae of *Pinirampus pirinampu* or *Brachyplatystoma* spp. were abundant in that period. The different sampling methods also presented expressive differences in the larval flux of some Pimelodidae species. *P. blochii, B. capapretum* and *B. rousseauxii* presented isolated peaks of larval flux detected by

The point sampling method carried out at fixed habitats along of transects is a good way to compare the larval abundance or composition in different depths or habitats [2, 3, 4, 5, 23]. However, when the aim is to assess the impact of the river modification over the ichthyo‐

(December-January) rising discharge.

0

0

cember-January) rising discharge.

only one method (Figure 6).

50

100

**Larval Flux**

150

200

50

100

**Larval Flux** 150

200

152 Biodiversity - The Dynamic Balance of the Planet

1 2 3 4 5 6 7 8 9 10 11 12 **Month**

> IJ IL PL

*Triportheus* **spp.**

1 2 3 4 5 6 7 8 9 10 11 12 **Month**

**Brycon spp.**

IJ IL PL

6).

**7. Discussion**

Figure 8. Monthly average of the larval flux (larvae/s) of the main species of Pimelodidae family: IJ- integrating sampling with juvenile net; IL- integrating sampling with larval net; and PL- point sampling with larval net. Months: 2- 5 (February-May) high discharge; 6-7 (June-July) falling discharge; 8-11 (August-November) low discharge; 12-1 (December-January) rising discharge. **Figure 8.** Monthly average of the larval flux (larvae/s) of the main species of Pimelodidae family: IJ- integrating sam‐ pling with juvenile net; IL- integrating sampling with larval net; and PL- point sampling with larval net. Months: 2-5 (February-May) high discharge; 6-7 (June-July) falling discharge; 8-11 (August-November) low discharge; 12-1 (De‐ cember-January) rising discharge.

plankton drifting process, it is necessary to develop a methodology adapted to estimate the larval flux in natural or dammed river cross sections. **7. Discussion**  The point sampling method carried out at fixed habitats along of transects is a good way to compare the larval abundance or composition in different depths or habitats [2, 3, 4, 5, 23]. However, when the aim is to

The three methods caught few eggs in relation to total of ichthyoplankton, and it was similar for the up and downriver sites. [4] also found few eggs in the ichthyoplankton composition in the Solimões River (Table 2). They hypothesized that the eggs have a short residence time, less than 16 hours, and as the spawning occurs mainly during the dusk, the eggs had hatched before assess the impact of the river modification over the ichthyoplankton drifting process, it is necessary to develop a methodology adapted to estimate the larval flux in natural or dammed river cross sections. The three methods caught few eggs in relation to total of ichthyoplankton, and it was similar for the up and downriver sites. [4] also found few eggs in the ichthyoplankton composition in the Solimões River

the sampling moment. The point sampling and the larval net caught more individuals than the integrating sampling and the juvenile net. However, the juvenile net caught mainly larvae in post-flexion stages and juveniles when compared to samples collected by larval net (Table 3 and 4). Using juvenile net is appropriated to sample ichthyoplankton of species that go through the study area in advanced development stages, like *Pimelodus blochii* and *B. rous‐ seauxii*.

in the abundance of Clupeiformes and Perciformes larvae in the Solimões River and Madeira River is difficult to be explained only by the methodology and the environmental difference of both rivers must be considered. The selectivity of the method for different species was very clear, while IJ method was very selective for Siluriformes larvae, in particular for the Pimelo‐ didae family; PL method was more selective for the Characiformes larvae, in particular for the Curimatidae and Auchenipteridae families (Table 6). For most families, larvae were found in pre-flexion or earlier stages, except for Characidae and Pimelodidae that showed larvae in

Diversity and Abundance of Fish Larvae Drifting in the Madeira River, Amazon Basin: Sampling Methods Comparison

http://dx.doi.org/10.5772/57404

155

The larvae drift indicated the reproductive period of the fish. The IL and PL methods identified a short annual reproductive period for Curimatidae and Prochilodontidae families, between December and April, and a prolonged annual reproductive period for Auchenipteridae and Anostomidae families, between November and May. However, the three methods showed an almost continuous reproductive period for Pimelodidae and Characidae families, with some

The larval flux estimated by IL and PL methods showed a short annual reproductive period, between December and January, for *Brycon* spp. and *Piaractus brachypomus,* and an annual prolonged reproductive period for *Mylossoma* spp., between December and March. Both methods indicated also a short annual reproductive period for *Triportheus* spp., but IL method indicated March as the reproductive period and PL method indicated May (Figure 8). This difference may be related with the selectivity of the different methods for the three species of *Triportheus* in the area. Thereby, it is necessary to identify the effect of the methods in the

The different methods presented peaks of reproduction activity in different months for *Brachyplatystoma* and *Pimelodus* species (Figure 8), which must be biased by the few number of larvae sampled. The IJ method collected mainly larvae of *P. blochii* and *B. rousseauxii,* and its larval flux showed a biannual short reproductive period for *P. blochii*, in March and in November, and a continuous reproductive period for *B. rousseauxii*, more intensive in the January and decreasing until December. *P. blochii* is considered a species that shows short annual reproductive period, which occurs in the beginning of rainy season [24]. The *P. blochii* larval flux peak occurred in March, at the end of the rainy season, and in November, at the beginning of the rainy season. As the studied area receives the flow discharge of the Beni and Mamoré Rivers, it must be investigated if the two larval peaks found in the area (Figure 8) have originated at the different basins upriver. *B. rousseauxii* reproduces in Andes foothill and the eggs and larvae drift the river toward to Amazon estuary [9]. Studies in Madre de Dios River in Peru, considered the main tributary of the Beni River, showed that this area is a spawning area for *B. rousseauxii* and the reproductive period is prolonged, and spawning period is concentrated at the high water period [5]. The larval flux peak of *B. rousseauxii* larval is in January (Figure 8) and the presence of larvae in advanced developmental stages is in accordance with the drifting movement from the Andes foothill during the high discharge. Finally, the impact assessment of the new infrastructures projects in the Amazon depends of the data quality obtained before the impact. The present study discussed the effect of the sampling method in the evaluation of the larval drift pattern in the Madeira River. The result

more advanced stages, such as *Brycon* spp., *Pimelodus blochii* and *B. rousseauxii*.

differences in the moment of the most intensive reproductive peak (Figure 6).

evaluation of the larval abundance of those species.

The abundance indexes, larval density and flux considered in this study, pointed an increasing of larvae and juveniles abundance during the rising discharge and the beginning of the high discharge phases (December to March), and a decreasing in the next months. It was also observed by [4], in Solimões River, and by [3], in the Madre de Dios River. However, the larval flux index was minimal from June to November while larval density index presented some expressive peaks during May to November. The relationship of both indexes was low, although had been statistically significant, for the IJ method and high for the IL and PL methods (Figure 5). The larval density is affected by the dilution effect, which occurs during the rising and high discharge, and by the concentrating effect, that occurs during the low water dis‐ charge. In that way, the larval density is a measure more accurate for the drift inversely proportional to the water discharge if the larval flux is constant. The larval flux is a more accurate measure to assess the changes of the larval drift in modified rivers. This measure is not biased by the discharge moment or by the width or depth of the river cross section.

The composition of the larval flux in relation to the larval development stages of each method (Table 4) indicated the IL method as the less selective in relation to the development stages of the fish when they are drifting in the river. The IJ method is selective for juveniles and underestimated the abundance of all larvae in early phases and the PL method is selective for larvae and underestimated the abundance of juvenile phase.

The larval flux estimated by the IL method was higher than the larval flux estimated by the PL method and they were similar in the upper and downriver sites, in spite of the number of larvae caught in the upriver was two times of the number caught downriver (Table 2 and 5). The two methods estimated the larval flux in pre-flexion stage as bigger than the larval flux in other stages and the larval flux in pre-flexion stage were more abundant in upriver than downriver, while the larval flux of the other stages were similar in both sites (Table 5). The upriver abundance of larvae in early stages may suggest a more intensive spawning activity more intensive upriver of the study area.

The identification of the larvae in rivers with high diversity is a challenge. Most larvae was identified at family level and it was possible to identify only few genera or species, mostly of the Characidae and Pimelodidae family. The larvae of Siluriformes and Characiformes orders as well as Curimatidae, Auchenipteridae, Pimelodidae, Characidae, Prochilodontidae, Anostomidae and Hemiodontidae families represented together 95% or more of all larvae caught during this study. These results do not corroborate the [4] results, in which the most abundant larvae found in the Solimões River were of the Clupeiformes, Characiformes and Perciformes orders. The small amount of larvae of Siluriformes order must be due to the differences in methodology. Most of the samples in the Solimões River were collected on the surface, while in the present study samples covered all water column. However, the difference in the abundance of Clupeiformes and Perciformes larvae in the Solimões River and Madeira River is difficult to be explained only by the methodology and the environmental difference of both rivers must be considered. The selectivity of the method for different species was very clear, while IJ method was very selective for Siluriformes larvae, in particular for the Pimelo‐ didae family; PL method was more selective for the Characiformes larvae, in particular for the Curimatidae and Auchenipteridae families (Table 6). For most families, larvae were found in pre-flexion or earlier stages, except for Characidae and Pimelodidae that showed larvae in more advanced stages, such as *Brycon* spp., *Pimelodus blochii* and *B. rousseauxii*.

the sampling moment. The point sampling and the larval net caught more individuals than the integrating sampling and the juvenile net. However, the juvenile net caught mainly larvae in post-flexion stages and juveniles when compared to samples collected by larval net (Table 3 and 4). Using juvenile net is appropriated to sample ichthyoplankton of species that go through the study area in advanced development stages, like *Pimelodus blochii* and *B. rous‐*

The abundance indexes, larval density and flux considered in this study, pointed an increasing of larvae and juveniles abundance during the rising discharge and the beginning of the high discharge phases (December to March), and a decreasing in the next months. It was also observed by [4], in Solimões River, and by [3], in the Madre de Dios River. However, the larval flux index was minimal from June to November while larval density index presented some expressive peaks during May to November. The relationship of both indexes was low, although had been statistically significant, for the IJ method and high for the IL and PL methods (Figure 5). The larval density is affected by the dilution effect, which occurs during the rising and high discharge, and by the concentrating effect, that occurs during the low water dis‐ charge. In that way, the larval density is a measure more accurate for the drift inversely proportional to the water discharge if the larval flux is constant. The larval flux is a more accurate measure to assess the changes of the larval drift in modified rivers. This measure is not biased by the discharge moment or by the width or depth of the river cross section.

The composition of the larval flux in relation to the larval development stages of each method (Table 4) indicated the IL method as the less selective in relation to the development stages of the fish when they are drifting in the river. The IJ method is selective for juveniles and underestimated the abundance of all larvae in early phases and the PL method is selective for

The larval flux estimated by the IL method was higher than the larval flux estimated by the PL method and they were similar in the upper and downriver sites, in spite of the number of larvae caught in the upriver was two times of the number caught downriver (Table 2 and 5). The two methods estimated the larval flux in pre-flexion stage as bigger than the larval flux in other stages and the larval flux in pre-flexion stage were more abundant in upriver than downriver, while the larval flux of the other stages were similar in both sites (Table 5). The upriver abundance of larvae in early stages may suggest a more intensive spawning activity

The identification of the larvae in rivers with high diversity is a challenge. Most larvae was identified at family level and it was possible to identify only few genera or species, mostly of the Characidae and Pimelodidae family. The larvae of Siluriformes and Characiformes orders as well as Curimatidae, Auchenipteridae, Pimelodidae, Characidae, Prochilodontidae, Anostomidae and Hemiodontidae families represented together 95% or more of all larvae caught during this study. These results do not corroborate the [4] results, in which the most abundant larvae found in the Solimões River were of the Clupeiformes, Characiformes and Perciformes orders. The small amount of larvae of Siluriformes order must be due to the differences in methodology. Most of the samples in the Solimões River were collected on the surface, while in the present study samples covered all water column. However, the difference

larvae and underestimated the abundance of juvenile phase.

more intensive upriver of the study area.

*seauxii*.

154 Biodiversity - The Dynamic Balance of the Planet

The larvae drift indicated the reproductive period of the fish. The IL and PL methods identified a short annual reproductive period for Curimatidae and Prochilodontidae families, between December and April, and a prolonged annual reproductive period for Auchenipteridae and Anostomidae families, between November and May. However, the three methods showed an almost continuous reproductive period for Pimelodidae and Characidae families, with some differences in the moment of the most intensive reproductive peak (Figure 6).

The larval flux estimated by IL and PL methods showed a short annual reproductive period, between December and January, for *Brycon* spp. and *Piaractus brachypomus,* and an annual prolonged reproductive period for *Mylossoma* spp., between December and March. Both methods indicated also a short annual reproductive period for *Triportheus* spp., but IL method indicated March as the reproductive period and PL method indicated May (Figure 8). This difference may be related with the selectivity of the different methods for the three species of *Triportheus* in the area. Thereby, it is necessary to identify the effect of the methods in the evaluation of the larval abundance of those species.

The different methods presented peaks of reproduction activity in different months for *Brachyplatystoma* and *Pimelodus* species (Figure 8), which must be biased by the few number of larvae sampled. The IJ method collected mainly larvae of *P. blochii* and *B. rousseauxii,* and its larval flux showed a biannual short reproductive period for *P. blochii*, in March and in November, and a continuous reproductive period for *B. rousseauxii*, more intensive in the January and decreasing until December. *P. blochii* is considered a species that shows short annual reproductive period, which occurs in the beginning of rainy season [24]. The *P. blochii* larval flux peak occurred in March, at the end of the rainy season, and in November, at the beginning of the rainy season. As the studied area receives the flow discharge of the Beni and Mamoré Rivers, it must be investigated if the two larval peaks found in the area (Figure 8) have originated at the different basins upriver. *B. rousseauxii* reproduces in Andes foothill and the eggs and larvae drift the river toward to Amazon estuary [9]. Studies in Madre de Dios River in Peru, considered the main tributary of the Beni River, showed that this area is a spawning area for *B. rousseauxii* and the reproductive period is prolonged, and spawning period is concentrated at the high water period [5]. The larval flux peak of *B. rousseauxii* larval is in January (Figure 8) and the presence of larvae in advanced developmental stages is in accordance with the drifting movement from the Andes foothill during the high discharge.

Finally, the impact assessment of the new infrastructures projects in the Amazon depends of the data quality obtained before the impact. The present study discussed the effect of the sampling method in the evaluation of the larval drift pattern in the Madeira River. The result is background knowledge for the future studies to assess the impact of the Jirau hydroelectric power plant to the larval drifting in the Madeira River.

[3] Cañas CM, Pine WE. Documentation of the temporal and spatial patterns of Pimelo‐ didae catfish spawning and larvae dispersion in the Madre de Dios River (Peru): in‐ sights for conservation in the Andean-Amazon headwaters. River Research

Diversity and Abundance of Fish Larvae Drifting in the Madeira River, Amazon Basin: Sampling Methods Comparison

http://dx.doi.org/10.5772/57404

157

[4] Araujo-Lima CARM., Oliveira C. Transport of larval fish in the Amazon. Journal of

[5] Cañas CM, Waylen PR. Modelling production of migratory catfish larvae (Pimelodi‐ dae) on the basis of regional hydroclimatology features of the Madre de Dios Basin in

[6] Goulding M, Carvalho ML. Life history and management of the tambaqui (Colosso‐ ma macropomum, Characidae): An important Amazonian food fish. Revista Brasi‐

[7] Carvalho JL, Merona B de. Estudos sobre dois peixes migratórios do baixo Tocantins, antes do fechamento da barragem de Tucuruí. Amazoniana 1986; 9(4):595-607.

[8] Ribeiro MCLB, Petrere Jr M. Fisheries ecology and management of the jaraqui (Sema‐ prochilodus taeniurus, S. insignis) in central Amazon. Regulated Rivers: Research

[9] Barthem R., Goulding M. The Catfish Connection: Ecology, Migration, and Conser‐

[10] Barthem RB, Goulding M. An Unexpected Ecosystem: The Amazon revealed by the fisheries. Amazon Conservation Association (ACA) - Missouri Botanical Garden

[11] Almeida O, Lorenzen K, McGrath D. The commercial fishing sector in the Regional Economy of the Brazilian Amazon. In: R. Welcomme and Peter. T. (Org.). Proceed‐ ings of the Second International Symposium on the Management of Large Rivers for Fisheries. 1 ed. Bangkok: FAO-Regional Office for Asia and the Pacific/RAP Pulica‐

[12] Isaac VJ, Almeida MC. El consumo de pescado en la Amazonía brasileña. COPES‐

[13] Humphries P, Lake PS., Fish larvae and the management of regulated rivers. Regu‐

[14] Agostinho AA., Pelicice FM, Gomes LC. Dams and the fish fauna of the Neotropical region: impacts and management related to diversity and fisheries. Brazilian Journal

[16] Goulding M, Barthem R, Ferreira EJG. The Smithsonian Atlas of the Amazon. Wash‐

[15] Tollefson J. A struggle for power. Nature 2011; 479: 160-161 doi: 10.1038/479160a.

vation of Amazon Predators. New York, Columbia University Press; 1997.

southeastern Peru. Hydrolical Processes 2011. DOI: 10.1002/hyp.8192.

Application 2011; 27:602-611.

Fish Biology1998. 53(Supplement A) 297-306

leira de Zoologia, São Paulo 1982; 1:107-133.

CAALC Documento Ocasional 2011; 13: 43.

Biology 2008; 68(4, Suppl.): 1119-1132.

ington: Smithsonian Institution; 2003.

lated Rivers: Research Management 2000; 16:421–432.

and Management 1990; 5: 195-215

Press. Lima, Peru; 2007.

tion2004; 2 15-24
