**5. Ichthyoplankton abundance**

when the discharge is minimal and the river beach is exposed, usually between August and November, while the high water level phase is when the discharge is high and the river floods the marginal areas, it occurs generally between February and May. Other months are consid‐ ered transition phases when the water level is rising or falling. The rising water level phase is when begins the rainy season and the river discharge starts to increase, between December and January, and the falling water level phase is when the river discharge decreases and the flood retreats, between June and July. The diary discharge median estimated for the study

Interannual variations of the hydrological cycle were observed during the period studied, with a short and intensive high discharge phase in 2010 (<160 days and >39 thousands m3

/s, which was used to separate the low and high discharge phases.

/s) (Figure 3).

/s) in

period was 15 thousand m3

142 Biodiversity - The Dynamic Balance of the Planet

Oct-2009

0

5,000

10,000

15,000

20,000

**Discharge (m3/**

 **s)**

25,000

30,000

35,000

40,000

45,000

Dec-2009

Feb-2010

May-2010

river of the study area (discharge- continuous line; water level- dotted line).

Jul-2010

Oct-2010

High High

Dec-2010

Mar-2011

**Figure 3.** The hydrological cycle of the Madeira River Basin obtained at the Hydrological Station of Abunã, at the up‐

The Madeira River is a muddy river that receives a large amount of sediments rich in mineral salt, which comes from Andean region. Due to this, the monthly averages of the conductivity and pH are in general higher than the other rivers of the Amazon Basin, varying from 69 to 131 µS/cm and 6 and 8, respectively. The conductivity and pH values were inversely related

Low Low

May-2011

Jul-2011

Oct-2011

Dec-2011

Mar-2012

High

May-2012

Aug-2012

78

80

82

84

86

88

90

**Water Level (m)**

92

94

96

98

100

relation to other years (>190 days and <36 thousands m3

During the three years, 4,148 individuals were collected by 432 samples of ichthyoplankton realized monthly in four transects, with an average of 9.5 samples by transect and method. The number of samples were similar for each combination of sampling methods and nets, however there were more samples in downriver (Jirau: 56%) than in upriver (Abunã: 44%). This difference in samples number was due to the river width, which is related to the number of sampling sites for integrating sampling method. A total of 21,665 larvae (99%) and 282 eggs (1%) were collected. The point sampling (PL) method was more efficient at collecting eggs than the integrating sampling methods. The PL collected 53% of the eggs, followed by the integrat‐ ing sampling method with juvenile net (IJ), that collected 29%, and the integrating sampling method with larvae net (IL), that caught the remains 17% of the eggs. The PL method also collected more larvae (61%) than the other methods and the IL method collected more larvae (34%) than the IJ method (5%). The number of eggs was similar for the both sites, and in the upriver (Abunã) showed twofold more larvae than in downriver (Jirau), even at downriver showing more samples (Table 2).

The abundance index considered the average of the larval density and the estimative of the larval flux in the cross section of four transects along the Madeira River. Peaks of larval density and flux were observed during the rising discharge and the beginning of the high discharge phases (December to March) and a decreasing in the next months. However, larval flux was minimal from June to November while larval density showed some peaks during this period

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**

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**

**Figure 4.** Monthly variation average of the larval flux (larvae/s) (□) and larval density (larvae/m<sup>3</sup>)(●) estimated for (A) integrating sampling with juvenile net-IJ; (B) integrating sampling with larval net-IL and (C) point sampling with larval net-PL. Months: 2-5 (February-May) high discharge; 6-7 (June-July) falling discharge; 8-11 (August-November) Low

0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08

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

145

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 **Density**

**Density**

**Density**

0 200 400 600 800 1,000 1,200 1,400 1,600 1,800

0 3,000 6,000 9,000 12,000 15,000 18,000

0 3,000 6,000 9,000 12,000 15,000 18,000

discharge; 12-1 (December-January) rising discharge.

**Larval Flux**

**Larval Flux**

**Larval Flux**

**C**

**B**

**A**


**Table 2.** Number of eggs, larvae, juveniles collected in up, downriver places, and the combination of sampling methods and nets: IJ: Integrating sampling with juvenile net; IL: integrating sampling with larval net; and PL: point sampling with larval net.

The development stage dominant in all samples was pre-flexion (66%), followed by flexion (19%) and post-flexion (5%). This stage composition was similar for the both sites, but different for the net types, where more than 95% of the ichthyoplankton caught by IL and PL were in flexion or early stages and 99% of the ichthyoplankton caught by IJ was in flexion or older stages. Considering each development stage, about 2/3 of all larvae in larval yolk or pre-flexion stages were caught by the PL method, whilst 49% of all larvae in post-flexion stages and 71% of juveniles were caught by the IJ method. Only nine small fish in adult stage were collected during this study (Table 3).


**Table 3.** Composition of the larvae collected considering the development stages and in relation to the up and downriver places and the combination of sampling methods and nets (IJ: integrating sampling with juvenile net; IL: integrating sampling with larval net; and PL: point sampling with larval net).

The abundance index considered the average of the larval density and the estimative of the larval flux in the cross section of four transects along the Madeira River. Peaks of larval density and flux were observed during the rising discharge and the beginning of the high discharge phases (December to March) and a decreasing in the next months. However, larval flux was minimal from June to November while larval density showed some peaks during this period

collected more larvae (61%) than the other methods and the IL method collected more larvae (34%) than the IJ method (5%). The number of eggs was similar for the both sites, and in the upriver (Abunã) showed twofold more larvae than in downriver (Jirau), even at downriver

49%

51%

**IL** 30 11% 2,979 14% 34%

**IL** 19 7% 4,316 20% 66%

**PL** 65 23% 9,369 43%

**PL** 85 30% 3,881 18%

The development stage dominant in all samples was pre-flexion (66%), followed by flexion (19%) and post-flexion (5%). This stage composition was similar for the both sites, but different for the net types, where more than 95% of the ichthyoplankton caught by IL and PL were in flexion or early stages and 99% of the ichthyoplankton caught by IJ was in flexion or older stages. Considering each development stage, about 2/3 of all larvae in larval yolk or pre-flexion stages were caught by the PL method, whilst 49% of all larvae in post-flexion stages and 71% of juveniles were caught by the IJ method. Only nine small fish in adult stage were collected

**Methods And Nets Sites**

**Table 3.** Composition of the larvae collected considering the development stages and in relation to the up and downriver places and the combination of sampling methods and nets (IJ: integrating sampling with juvenile net; IL:

integrating sampling with larval net; and PL: point sampling with larval net).

**IJ IL PL Upriver Downriver Unknown** 22 403 318 423 320 743 3% **Larval Yolk** 7 376 790 910 263 1,173 5% **Pre-Flexion,** 7 4,599 9,713 9,830 4,489 141319 66% **Flexion** 396 1,667 2,073 2,368 1,768 4,136 19% **Post-Flexion** 522 220 319 647 414 1,061 5% **Juvenile** 159 30 35 134 90 224 1% **Adult** 7 0 2 4 5 9 0% **Total** 1,120 7,295 13,250 14,316 7,349 21,665 100%

631 3%

489 2%

**Total**

**Sites S&N Eggs Larvae & Juveniles**

**Total** 282 100% 21,665 100%

**Table 2.** Number of eggs, larvae, juveniles collected in up, downriver places, and the combination of sampling methods and nets: IJ: Integrating sampling with juvenile net; IL: integrating sampling with larval net; and PL: point

**IJ** 53 19%

**IJ** 30 11%

showing more samples (Table 2).

144 Biodiversity - The Dynamic Balance of the Planet

**Upriver (Abunã)**

sampling with larval net.

during this study (Table 3).

**Stage**

**Downriver (Jirau)**

**Figure 4.** Monthly variation average of the larval flux (larvae/s) (□) and larval density (larvae/m<sup>3</sup>)(●) estimated for (A) integrating sampling with juvenile net-IJ; (B) integrating sampling with larval net-IL and (C) point sampling with larval net-PL. 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 4). The relationship between the two abundance indexes for the three methods is presented in Figure 5. The highest correlation value between larval density and flux was observed for the IL method (r2 =0.86) and the lowest values was observed for the IJ method (r2 =0.23). The correlation between IL and PL methods was higher for larval flux (r2 =0.77) than for larval density (r2 =0.61) (Figure 5). The mean composition of the larval flux by larval stages of 144 samples shows the importance of IJ and IL methods for the juvenile abundance estimative and the IL and PL for the abundance estimative of the early stages (Table 4).

Levene's test (F=0.31, p>0.05). The larval flux estimated by PL method is significantly lower than the IL method for all stages (Method: F=7.53, p<0.01), especially for juvenile stage, but the interactive effects of sampling method with the other factors was not significant (Method-Site, Method-Stage, Method-Site-Stage: p>0.05). The larvae of the pre-flexion stage were more abundant than the other stages (Stage: F=26.10, p<0.01). The up and down river sites was not a significant factor (Site: p>0.05), but the interactive effect with development stage was significant (Site-Stage: p<0.05) indicating larvae in pre-flexion stage is higher in upriver than downriver, but the other stages presented similar values in both sites (Tukey HSD test, p<0.01)

**Table 4.** Average and percent of the development stage composition for each method considering larval flux (larvae/s). Methods: IJ- integrating sampling with juvenile net; IL- integrating sampling with larval net; and PL-point

**Stage IJ IL PL Total**

**Larval Yolk** 12 (1%) 685 (55%) 539 (44%) 1,236 (100%) **Pre-Flexion** 13 (0.2)% 3,675 (48%) 3,996 (52%) 7,684 (100%) **Flexion** 217 (7%) 1,507 (52%) 1,175 (41%) 2,898 (100%) **Post-Flexion** 344 (19%) 677 (38%) 758 (43%) 1,779 (100%) **Juvenile** 198 (41%) 201 (42%) 82 (17%) 481 (100%) **Total** 822 (5%) 7,261 (48%) 6,944 (46%) 15,027 (100%)

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

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

147

*Intercept* 580306.0 1 580306.0 689.5715 <0.0001\*\* *Method* 6340.9 1 6340.9 7.5349 0.0062\*\* *Site* 1882.5 1 1882.5 2.2369 0.1352 *Stage* 87861.8 4 21965.5 26.1013 <0.0001\*\* *Method-Site* 23.4 1 23.4 0.0278 0.8676 *Method-Stage* 891.3 4 222.8 0.2648 0.9006 *Site-Stage* 8308.0 4 2077.0 2.4681 0.0436\* *Method-Site-Stage* 2193.1 4 548.3 0.6515 0.6260

**Table 5.** Result of Factorial ANOVA testing the methods (IL and PL), river site (up and downriver) and development stage (larval yolk, pre-flexion, flexion, post-flexion and juvenile) factors over the square root of the larval flux. S.S. = sum of squares; d.f. = degrees of freedom; M.S. = mean square; F = F statistic; p = significance level. One asterisk means

*Error* 637050.2 757 841.5

significant at 0.05 level and two asterisks means significant at 0.01.

*SS d.f. MS F p*

(Table 5).

sampling with larval net.

**Figure 5.** Abundance indexes plots for the three sampling methods and the respective Person's coefficient *r* (\* p<0.01, n=144; n.s: no significant). Abundance indexes: Density (larvae/m3) and Flux (larvae/s). Methods: IJ: integrating sam‐ pling with juvenile net; IL: integrating sampling with larval net; and PL: point sampling with larval net.

The seasonal and spatial variation of the larval flux was analyzed only for IL and PL data. The IJ data was not considered due its low capacity to detected larvae in early stages and to assure the assumptions for homogeneity of variance. An ANOVA-two-way was performed to analyze the variation in larval composition considering the square root of the larval flux as dependent variable and the sampling method (IL and PL), the up and downriver sites and five develop‐ ment stages (larval yolk, pre-flexion, flexion, post-flexion and juvenile) as independent variables (Table 5). The assumptions for homogeneity of variance was met according to the Diversity and Abundance of Fish Larvae Drifting in the Madeira River, Amazon Basin: Sampling Methods Comparison http://dx.doi.org/10.5772/57404 147


(Figure 4). The relationship between the two abundance indexes for the three methods is presented in Figure 5. The highest correlation value between larval density and flux was

of 144 samples shows the importance of IJ and IL methods for the juvenile abundance estimative and the IL and PL for the abundance estimative of the early stages (Table 4).

=0.23). The correlation between IL and PL methods was higher for larval flux (r2

Density- PL

0.16 n.s. 0.76\*

0.41\*

=0.86) and the lowest values was observed for the IJ method

=0.61) (Figure 5). The mean composition of the larval flux by larval stages

0.48\*

0.16 n.s. 0.05 n.s.

Flux- IJ

0.53\* 0.91\*

0.47\*

0.92\* 0.78\*

0.41\*

0.55\*

**Figure 5.** Abundance indexes plots for the three sampling methods and the respective Person's coefficient *r* (\* p<0.01, n=144; n.s: no significant). Abundance indexes: Density (larvae/m3) and Flux (larvae/s). Methods: IJ: integrating sam‐

The seasonal and spatial variation of the larval flux was analyzed only for IL and PL data. The IJ data was not considered due its low capacity to detected larvae in early stages and to assure the assumptions for homogeneity of variance. An ANOVA-two-way was performed to analyze the variation in larval composition considering the square root of the larval flux as dependent variable and the sampling method (IL and PL), the up and downriver sites and five develop‐ ment stages (larval yolk, pre-flexion, flexion, post-flexion and juvenile) as independent variables (Table 5). The assumptions for homogeneity of variance was met according to the

pling with juvenile net; IL: integrating sampling with larval net; and PL: point sampling with larval net.

Flux- IL

Flux - PL

0.53\* 0.55\*

0.91\*

=0.77) than

0.05 n.s.

0.79\*

0.85\*

observed for the IL method (r2

146 Biodiversity - The Dynamic Balance of the Planet

Density- IL

0.78\*

0.92\*

0.79\*

0.16 n.s.

0.47\*

0.76\*

0.85\*

for larval density (r2

Density- IJ

0.16 n.s.

0.48\*

0.05 n.s.

0.05 n.s.

(r2

**Table 4.** Average and percent of the development stage composition for each method considering larval flux (larvae/s). Methods: IJ- integrating sampling with juvenile net; IL- integrating sampling with larval net; and PL-point sampling with larval net.

Levene's test (F=0.31, p>0.05). The larval flux estimated by PL method is significantly lower than the IL method for all stages (Method: F=7.53, p<0.01), especially for juvenile stage, but the interactive effects of sampling method with the other factors was not significant (Method-Site, Method-Stage, Method-Site-Stage: p>0.05). The larvae of the pre-flexion stage were more abundant than the other stages (Stage: F=26.10, p<0.01). The up and down river sites was not a significant factor (Site: p>0.05), but the interactive effect with development stage was significant (Site-Stage: p<0.05) indicating larvae in pre-flexion stage is higher in upriver than downriver, but the other stages presented similar values in both sites (Tukey HSD test, p<0.01) (Table 5).


**Table 5.** Result of Factorial ANOVA testing the methods (IL and PL), river site (up and downriver) and development stage (larval yolk, pre-flexion, flexion, post-flexion and juvenile) factors over the square root of the larval flux. S.S. = sum of squares; d.f. = degrees of freedom; M.S. = mean square; F = F statistic; p = significance level. One asterisk means significant at 0.05 level and two asterisks means significant at 0.01.
