Odonatologica 39(4): 287-303 December /, 2010

Richness and structure of an Odonata larval

assemblage from Rio Pinolapa, Tepalcatepec,

Michoacán, Mexico in relation to their habitat

characteristics

J.A. Gómez-Anaya and R. Novelo-Gutiérrez

Institute de Ecología, A.C., Apartado Postal 63,

MX-91070, Xalapa, Veracruz, Mexico

[email protected],mx; — [email protected]

Received January 7, 2010 / Reviewed and Accepted April 4, 2010

The odon. larval assemblage from Rio Pinolapa (RP) in the municipality of Tepal-
is described.
catepec, Michoacan, Sampling conducted twice in each (8
was season

in and variables of the river channel

trips total), additionally some physicochemical

were recorded. Strata (shores, riffles and eddies) and seasonal variation of
assemblag-
described
es are and compared using classical diversity measures such as Shannon’s

diversity index, Simpson’s diversity index as a dominance measure, Margalef’s rich-

ness index and Pielou’s evenness index. For comparing strata and seasonal diversity
the Renyi’s diversity profiles were used. A Cluster Analysis was performed on a Bray-

-Curtis similarity matrix to explore the faunal relationships seasons and

among year

strata. CCA was also performed to investigate the relationships between the physi-
cochemical and species abundance matrixes. As results, 28 spp. (12 Zygoptera and

16 Anisoptera) were recorded as larvae. Most abundant species were Erpetogomphus

elaps, Brechmorhogapraecox and Phyllogomphoides luisi. The highest number of spp.

registered in winter and the lowest in strata the abun-

was summer.
Among highest
dance was recorded in riffles, although the shoreline had the largest number of spp.

The most similar assemblages were those of autumn and winter. Shore habitats were

more heterogeneous than eddies and riffles and this could explain the larger number

of species. The Clench’s model explains better the data. Additionally, we used the

slope of cumulative number of curve for assessing completeness of the RP list.

spp.

CCA was
significant, with pH, autumn, shoreline and riffles the most important vari-

ables. This means that species variation is related to physicochemical, temporal and

strata conditions in RP.

288 J.A. Gomez-Anaya & R. Novelo-Gutierrez

INTRODUCTION

The estimation of biodiversity has become one of the major goals for ecolo-

gists up to now (MAGURRAN, 2004). Conservation of biodiversity requires the

knowledge of its patterns and magnitude (BASELGA & NOVOA, 2008), where

the process of making an inventory of unknown assemblages, as well as describ-

ing new species, represents the first step in understanding that biodiversity, and

also the foundation for later research related to biodiversity. Unfortunately,

any

this kind of work has received relatively little attention by many journals, propi-

et
tiating a great abandonment by researchers (WHEELER, 2004; WHEELER

al., 2004; DE CARVALHO et al„ 2005).

Different methods for assessment have been in-
biodiversity developed as an

dicator of ecosystem conditions, conservation goals, management and environ-

mental monitoring (SPELLERBERG, 1991). Usually, the number of species

is the more
used measure
of diversity (MAYR, 1992; MORENO. 2000). How-

diversity depends not only on

the number of species but also on
the relative
ever,

abundance of them. Generally, species are distributed according to hierarchical

abundance classes, from some very

abundant to some
very rare.
As some species
become more abundant and others become rarer, the lower the biodiversity of

the community will be. Then, the conservation of biodiversity is mainly a prob-
lem concerning to the ecological behavior of the rare species.
In this manner, measuring the relative abundance of species will allow us to

identify those species that, due to their rareness in the

community, are more sen-

sitive to environmental changes (MORENO, 2001) and, consequently, more ex-

posed to extirpation.
On the other hand, Odonata have become among the most used aquatic insect

groups in ecological quality assessment today, because they are relatively large,

abundant, widely distributed, easy to collect, the larvae are rather sedentary and

to rear in laboratory, and there is limited genetic variation (HELLAWELL,

easy

1986). Also, they are relatively easy to identify in comparison with other groups

and well in aquatic samples (HAM-

(STORK, 1994), are generally represented

MOND, 1994). Moreover, they respond quickly to environmental stress (NOSS,

1990). Fortunately, the identification of immature stages has been developed

in the last two decades in Mexico. However, an important gap
still remains on

the knowledge of the structure and function of Mexican Odonata assemblages,

with relatively few studies dealing with this subject (NOVELO-GUTIERREZ
& GONZALEZ-SORIANO, 1991; GOMEZ-ANAYA et al.. 2000; NOVELO-

-GUTIERREZ et al., 2002; ALONSO-EGUIALIS, 2004). The exploration of

patterns in time and space of Odonata larval assemblages can supply basic data

for future research.

The goal of this work is to describe the structure and seasonal variationof the

Odonata larval assemblage from the Rio Pinolapa and relate it to environmental

factors.
 Odonata larval assemblage from Rio Pinolapa 289

STUDY AREA

The Rio Pinolapa (RP) is located at (19°00.524N; 103°01.456W), municipality of Tepalcatepec,

in Michoacan State, Mexico (Fig. 1). The sampling site is at 616 m asl. Average gradient was 0.02

=
(I °08’44.75”,n = 7), minimum = 0.006 (0°20’37.57”),maximum 0.042 (2°24’ 18.03”). Average depth
=
0.11 m
(IC =
0.05 -
0.19 m, n =
8); average width
=
2.18 m
(IC =
0.65 -
3.71 m, n = 8); current ve-

3 5
m/s, 9.16m /s(IC 3.32-21.65 Aver-
locity =
37.58 m/s(IC= 28.57-46.58 n =
8); discharge = = m /s).
of physicochemical variables were: temperature
=
28.03°C (IC =
27.02- 29.04°C, n = 30), pH =
ages

8.47 (IC = 8,33-8.61, n = 30), conductivity = 666.83 pS/cm (IC = 640.70 -

692,96 pS/cm, n = 30) and

= 7.78 (1C = 7.05 8.51 n = 30).

oxygen ppm -

ppm,

site in of Mexico.
Fig. 1. Sampling (empty circle) municipality Tepalcatepec, Michoacan,

MATERIAL AND METHODS

COLLECTING. —

Larvae were collected twice in each season (8 trips in total) at shores, and in

riffles (at mid-channel) and eddies. Usually, sampling was done at the end of the first third and at the

beginning of the last third of each season. We used an aquatic D-frarae net, and samples were
pre-

served in 96% alcohol with replacement before 24 h. A stereomicroscope used to

one was
separate
and quantify all larvae to the species level. In addition to the larval sampling, physicochemical vari-

ables such dissolved

as pH, oxygen, temperature and conductivity were recorded for each sample.

Depth, width, and current velocity were measured and discharge was then calculated. The gradient
(slope) was measured at seven points of the 500 m long sampling transect according to RESH et al.

(1996).

DIVERSITY MEASURES AND DATA ANALYSIS. -

Richness and composition, as well as

classical such the Shannon’s diversity index the Simpson’s index ( D), the
diversity measures as (//').

richness index ( and the Pielou’s index used in order to describe the
Margalef’s R), evenness
(J) were

Odonata assemblages by season, strata and as a whole (MORENO, 2001). Also, the Renyi’s diversity

profiles (TOTHMERESZ, 1995,1998; JAKAB, 2002) were used for comparing diversity, as proposed
by SOUTHWOOD & HENDERSON (2000). In this method when the value of the scale used as a

parameter is low, the method is extremely sensitive to the

presence
of rare species. As the value of the

scale increases, diversity is less sensitive to rare species. At a high value, the method is sensitive only

The result of this scale-dependent characterization of diversity be used in

to common species. can

form to visualize the relations of This is usually called ‘the

a
graphical diversity assemblages. curve
290 J.A. Gomez-Anaya & R. Novelo-Gutierrez

diversity profile of the assemblage’. It is important to stress that curves of two diversity profiles may

intersect. For two communities, the intersection of the diversity profiles means that one of the com-

munities is more diverse for rare species, while the other oneis more diverse for common species. The

Species Diversity & Richness package 3.0 employed to the Renyi’s values, exporting
v. was
generate
them to an Excel spreadsheet to show them graphically.
THEORETICAL RICHNESS. —
An estimate of the theoretical richness using non paramet-

ric estimators Chao2, Bootstrap, and upper limit of Mao Tau was carried out, using Estimates 8.0

(COLWELL, 2006). Additionally, parametric methods as richness estimators by extrapolation were

also which the observed of species accumulation for

used, apply curve
modeling the addition of new

in relation to the sampling effort (PALMER, 1990; SOBERON

species & LLORENTE, 1993). The
Clench’s (CLENCH, 1979) and Linear dependence models were applied, as exemplified by JIMEN-

EZ-VALVERDE & HORTAL (2000). Likewise, the slope on the cumulative species curve was used

to assess the completeness of assemblages (HORTAL & LOBO, 2005). The slopes were obtained by

means of the first derivative of the Clench’s and Linear dependencefunctions (NOVELO-GUT1ER-

REZ & GOMEZ-ANAYA, 2009).

CLUSTER ANALYSIS. — Beta diversity was assessed by methods of classification. A Cluster

Analysis (CA) on a
Bray-Curtis (BC) similarity matrix [(1-W) where W =
BC dissimilarity)] and the

Unweighted Pair Group Method with Arithmetic mean

(UPGMA) were used to explore the faunal

relationships seasons and strata. This analysis performed using PC-ORD 4.5
among was ver
(Mc-
CUNE & GRACE. 2002).

CANONICAL CORRESPONDENCE ANALYSIS. -

The Canonical Correspondence Anal-

ysis (CCA), a direct ordination method, was used to relate species abundance with environmental

variables (TER BRAAK & SMILAUER, 1998). The number of environmental variables was then

reduced using the automatic forward selection option in the CANOCO 4.5 The statistical
program.

significance of the relationship between the species and the set of environmental variables was tested

by a Monte Carlo permutation test, using an F-ratio of the sum of all eigenvalues as the statistical

test (TER BRAAK and PRENTICE, 1988).

RESULTS

SEASONAL PHYSICOCHEMICAL VARIATION

Temperature, pH and conductivity had higher averages in summer, while ox-

ygen was highest in spring. In fact, oxygen decreases gradually from spring to

winter (Tab. I). All pH values were slightly alkaline. Oxygen levels were very low

in winter when both abundance and species richness were highest.

LARVAL RICHNESS AND COMPOSITION

A total of 3,278 Odonata larvae belonging to 28 species (12 Zygoptera and 16

Anisoptera), 16 genera and six families were collected (Tab. II). Erpetogomphus

elaps (50.21%) was the dominant species; other numerically important species

were Brechmorhoga praecox (14.16%) and Phyllogomphoides luisi (6.72%) (Fig.

2). A further 61.54% of all species occurred in low abundance (<1%) and were

considered rare.
 Odonata larval assemblage from Rio Pinolapa 291

SEASONAL NUMERICAL DOMINANCE

95% 11.45 8.23 9.36 4,69

During spring Erpetogomphus elaps (38.73%), Argia oe-

nea (17.40%), and Progomphus marcelae (14.95%) domi-

-95% 9.55 6.33 8.01 2.78 nated numerically. Only seven species were detected in

summer (four of which were gomphids), with low

very

total and relative abundances. No coenagrionid was re-

Oxygen 10.5 7.28 8,68 3.74 corded in this season. E. elaps (44.15%) and B.
praecox

(28.39%) dominated in autumn. Finally, during winter

E. daps (60.10%) dominatedthe assemblage. Argia tezpi

Pinolapa 95% 765.1 810.7 609.4 798.9 (8.74%)

as
and Phyllogomphoides.
codominant species
together
luisi (6.52%)

with E.
were

daps.
present

It is in-

Rio teresting to note that while E.

elaps clearly dominated

of
-95% 647.6 693.3 526.3 681.4 throughout the year in the larval
stage,
the imagoes were

not
very commonly encountered.

vari bles
Conductivy
SEASONAL ASSEMBLAGES

706.3 752.0 567,83 740.17

physichemical Table III and

rameters.
Figure 3 show the seasonal

The smallest number of

species
ecological
was recorded
pa-

I in spring and autumn were intermediate, and

summer,
the 95% 8.81 9.08 8.19 8.63
Table for
the highest number of species was
found in winter. The

abundance pattern seems to follow the richness one,

be-

intervals -95% 8.56 8.83 8.02 8.37

ing

diversity
higher in winter and lower in

index (//’) seems to vary little

summer.
Shannon’s

throughout the

four seasons, although it was a little higher in spring.

confidenc PH 8,68 8.95 8.11 8.50

Dominance ( D
) was higher in winter, mainly

great abundance of K elaps. Renyi's diversity profiles

due to the

are

shown in Figure 4. The summer pattern was a straight

95%
line. It showed the minimum number of species for a =

and 95% 32.51 35.56 26.61 21,54 0 (the basic structure of assemblages), but for values
up

2 (Simpson index) the pattern showed that

Averages
to summer

diversity was the highest.

-95% 31.29 34.34 25.75 20.32
SIMILARITY

Tempratue
The Odonata larval assemblages from autumn and
31.90 34.95 26.18 20.93 winter the due their
were most similar, mainly to high
and similar abundance (Fig. 5), nevertheless, they were

quite different in richness sharing a high number of spe-

cies Some like
(13). species Argia oculata, Erpetogom-

Spring Sum er Aut mn Winter phus cophias and Paltothemis lineatipes were exclusively
292 J.A. Gomez-Anaya & R. Novelo-Gutierrez

Table II

Richness and composition of seasonal Odonata larval assemblages from Rio Pinolapa

Taxa Spring %
% Summer
Summer %
% Autumn
Autumn %
% Winter
Winter %
% Total
Total %
%
Key
Key Spring

Number
Number of individuals
individuals 410 12.45
12.45 20
20 0.61 1452
1452 44.32
44.32 1396 42.61
42.61 3278
3278

Number
Number of species
species 15
15 55.56 7 25.93 16 59.26 22
22 81.48 100

ZYGOPTERA

Calopterygidae
Hetaerina
Hetaerina

americana
americana Heam
Hearn - - 4 20.0
20.0 22 1.52 7 0.50 33 1.01

Platystictidae
Platystictidae
Palaemnema

domina Pado 23
23 5.64
5.64 - 16
16 1.10 54 3.87 93 2.84

Coenagrionidae
Coenagrionidae
Argia funcki
Argiafuncki Arfu 7
1 M2
1.72 -. . -. . _ _ 7 0.21

A. oculata
A. Aroc
A 3
3 0.21 3
3 0.09
- -

roc -- - - - -

A.
A. Aroe
roe 71 17.40 14
14 0.96 53
53 3.80
3.80 138
138 4.21
4.21
oenea - -

A. pallens
liens Arpa 1 0.25 -
8
8 0.57 9 0.27
pa
-- - --

A. pulla 0.43 6 0.18

pulla Arpu 6 0.43 6 0.18
- - - - -- -

A. tezpi
A. Arte 1 0.25
0.25 20 1.38 122 8.74
8.74 143
143 4.37
tezpi -- -

Enallagma
Enallagma

novaehispaniae
novaehispaniae Enno 5 1.23
1.23 -. . -. . 2
2 0.14 7 0.21
0.21

E. semicirculare Ense - -- -- - -- - 3
3 0.21 3 0.09
0.09

Telebasis salva
Telehasis Tesa 2
2 0.49 - - - - - 2 0.06

Protoneuridae

Protoneura cara Prca - -

-
1
1 0.07
0.07 11 0.03
0.03

ANISOPTERA
ANISOPTERA

Gomphidae

Erpetogomphus

bothrops
hothrops Erbo
Erho - - - 1
1 0.07 11 0.03
0.03

E. cophias Erco
Erco - - - I1 0.07 - - 11 0.03

E. elaps Erel 158 38.73 7 35.0

35.0 641 44.15 839
839 60.10 1645
1645 50.21

Progomphus
clendoni Prcl 13
13 3.19
3.19 2 10.0 9 0.62 8 0.57 32 0.98

lambertoi
P. lambertoi
P Prla
Prla 11 0.25
0.25 - - 91
91 6.27
6.27 11 0.07
0.07 93
93 2.84
2.84

P. marcelae
P Prma
Prtna 61
61 14.95 - - 8 0.55 65
65 4.66 134 4.09

Phyllogomphoides
s

luisi Phlu 21 5.15

5.15 2 10.0 106 7.30 91
91 6.52 220 6.72

P. pacificus
P. Phpa - - 2 10.0 9 0.62 8 0.57 19 0.58

Libellulidae
Libellulidae

Dythemis
7
7 0.50
0.50 7
7 0.21
0.21
nigrescens Dyni
Dyni - - - - -

Brechmorhoga
Brechmorhoga
5
5 1.23 2
2 10.0 412
412 28.37
28.37 45 3.22
3.22 464 14.16
praecox Brpr
Brpr

Erythrodiplax 0.00
0.00 7
7 0.50
0.50 7
7 0.21
0.21
Erythrodiplax sp. Ersp
-- - - -
,

Macrothemis
Macrothemis

inacuta Main
Main 1
7 1.72
1.72 - - - 7 0.21
0.21

M.
M. pseudimitans 28 6.86 1
1 5.0
5.0 83 5.72
5.72 60
60 4.30 172
172 5.25
Maps

Paltothemis

lineatipes
lineatipes Pali - - - - 3 0.21
0.21 - - 3 0.09
0.09

Perithemis

domitia
domitia Pedm - - - - - 2 0.14 2 0.06
0.06

Pseudoleon

Pssu
Pssu 6 1.47
1.47 - 14
14 0.96
0.96 6
6 0.43
0.43 26
26 0.79
0.79
superbus
superhus
-
 Odonata larval assemblage from Rio Pinolapa 293

Fig. 2. Relative abundance of Odonata species from Rio Pinolapa. Key to species in Table II.

recorded in autumn, while others as

Argia pulla, Enallagma semicirculare, Pro-

toneura cara, Dythemis nigrescens and Erythrodiplax sp., were only found in win-

ter. Summer contains an assemblage of very few species.

DIVERSITY AND ABUNDANCE BY STRATA

Table IV shows the ecological parameters of the Odonata larval assemblages

by strata of RP. Number of species in the shores was

almost twice as
much of

that of the riffles (middle-channel), and more than twice as

much of that of ed-

dies. However, the abundance was higher in the middle-channelof the water body.
The diversity H’ was higher in the shores while the dominance did in eddies.

The major amount of larvae in all strata was Erpetogomphus elaps, however,
in the riffles found higher proportion of
we a Brechmorhoga praecox (24.42%).
When diversity is compared and ordered using the Renyi’s profiles (Fig. 6), it

Table III

Seasonality of the ecological parameters of Odonata larval at Rio

assemblages Pinolapa

Spring Summer Autumn Winter Total

Number of species 15 7 16 22 28

Number of specimens 401 20 1452 1396 3269

Simpson (D) 0.22 0.21 0.29 0,38 0.29

Shannon-Weaver (//’) 1.89 1.76 1.63 1.61 1.85

Margalef richness ( R) 2.34 2.00 2.06 2.90 3.09

Pielou evenness (J) 0.70 0.90 0.59 0.52 0.57

294 J.A. Gomez-Anaya & R. Novelo-Gutierrez

Fig. 3. Ecological parameters per season of the Odonata larval assemblage at Rio Pinolapa. Collec-

tions were made between March 2005 and January 2006.

follows the gradient shores>riffles>eddies. This fact confirms that shores contain

the highest diversity of Odonata larvae in RP.

THEORETICAL RICHNESS ESTIMATION

NON-PARAMETRIC MODELS. —

The cumulative species curves generated by

non parametric estimators Chao2, and Bootstrap are

shown in Figure 7. The es-

timated number of species was 41.4, and 32.4 species, respectively, which
gave a

sampling efficiency of 67.6%, and 86.3, respectively. These richness estimators

indicate a lack of register from 4 to 13 species. The estimated number of species

using the Mao Tau upper limit of class interval was 34.7, which means that still

should be added to the list 6-7 species, being the efficiency of the total sampling
effort of 80%. The number of species with a single individual (singletons) was 3,

with individuals (doubletons) 2, the number of 11,

two was unique species was

and for duplicated ones was 3.

PARAMETRIC MODELS. -

Table IV
Figure 8 shows cumulative
Ecological parameters of the Odonata larval assemblages
species curves generated by by stratum at Rio Pinolapa
the Clench’s and Linear de-

pendence functions. The first Index/stratum Shores Riffles Eddies

function predicted 30.67 spe-

Number of species 25 14 10
cies and explained 98% of
Number of specimens* 1062 1713 323
data variation, while the sec-

ond one predicted 25.74 spe-

Simpson ( D) 0.27 0,32 0,47

cies and explained 95% of Shannon-Weaver (//’) 1.93 1.54 1.18

data variation. The Clench’s Margalef’s richness ( R ) 3.44 1.61 1.56

Pielou’s (J) eveness 0.60 0.60 0.51

model indicates that there are

3 species to be registered yet, *

Some samples were omitted because of insufficient field

and the linear in-

dependence data.
 Odonata larval assemblage from Rio Pinolapa 295

Fig. 4. Renyi’s diversity profiles for the four seasonal Odonata larval assemblages from Rio Pinola-

pa.

2
dicates the list has been
completed. Based on the explained variance ( R ,
deter-

minationcoefficient), the Clench’s estimation explained better the data variation;

its prediction is considered further. Finally, slopes for both curves were 0.10 for

Clench’s function and 0.04 for Linear dependence.

THE SPECIES-ENVIRONMENT RELATIONSHIPS

The results of the CCA were globally significant (trace =

0.937, F =
1.59, p

<0.05, Tab. V). The first three axes offered a good solution to the ordinationof

the physicochemical variables and abundanceof species, since from the total vari-

ability in the data (inertia =

3.701), it was possible to explain 87.9% by means
of

these group of axes. The significance test of the first canonical axis showed it was

significant (eigenvalue =
0.292, F =
2.822, p <0.05).

Fig. 5. Dendrogram showing the faunal relationships the seasons assemblages of odo-
among year

nate larvae. Based on a Bray-Curtis similarity matrix and the unweighted pair-group arithmetic av-

eraging (TJPGMA).
296 J.A. Gomez-Anaya & R. Novelo-Gutierrez

Table V

Results of the canonical correspondence analysis (CCA) of Odonata larvae

log-transformed
abundance as a function of their environmental variables

Axes 1 2 3 4 Total inertia

Eigenvalues 0.29 0.25 0.18 0.09 3.705

Species-environment correlations 0.82 0.85 0.76 0.66

Cumulative variance
percentage
of species data 7.9 14.7 19.5 22

of species-environment relation 31.2 58 76.9 86.8

Sum of all eigenvalues 3,705

Sum of all canonical eigenvalues 0.937

The first axis was the most important, explaining 31.2% of variance, and it was

also the most strongly correlated with pH, riffles and autumn. The second axis

explained 26.8% of variance, and it correlated strongly with conductivity. The

third and fourth axes explained only 16.7% and 12.1% of variance, respectively,

and were not

considered further.

When the distribution of the species in the eight collections and three strata

is analyzed together with the CCA of the Figure 9, it is posible to make the fol-

lowing precisions: Paltothemis lineatipes, Erpetogomphus cophias, Progomphus

Fig. 6. Renyi’s diversity profiles for three strata assemblages of Odonata larvae from Rio Pinolapa.
Profiles differ mainly at their basic level of structure, the number of species. Profiles never cross in

range from 1 to 4. Values of Renyi when 4 shores profile 0.982, riffles profile 0.969
=
a a = were: =

and eddies profile =

0.535.
 Odonata larval assemblage from Rio Pinolapa 297

lambertoi, and Hetaerina americana do associate well with the autumn. Particu-

larly, P. lambertoi was registered almost in 100% (91 out of 92 specimens) from

the shores at the beginning of the autumn. Argia oenea and Palaemnema domina

were almost invariably registered from the riffles and at late winter. Only seven

larvae of Macrothemis registered from late spring shores.

inequiunguis were on

DISCUSSION

GENERAL ASPECTS

A total of 28 Odonata and 16 Anisoptera) found

species (12 Zygoptera were

and 3,276 larvae were identified in this work. The size of the Odonata larval as-

semblage fromRP is similarto other Mexican water bodies reported (ALONSO-

EGUIALIS, 2004; GOMEZ-ANAYA et al„ 2000; NOVELO-GUTIERREZ &

GONZALEZ-SORIANO, 1991; NOVELO-GUTIERREZ et al„ 2002; BOND

et ah, 2006). Usually, the most speciose families are Libellulidaeand Coenagrio-
nidae. In this case Coenagrionidae, Gomphidae and Libellulidae were best rep-

resented. The single genus Argia contributed 50% of Coenagrionidae; in Gom-

phidae three contributed more or

less equally, and in Libellulidae seven
genera

contributed to the diversity.

genera

DIVERSITY AND ABUNDANCE

The highest and lowest species richness was found in winter and summer, re-

spectively. The abundance of individuals recorded in autumn and

greatest was

Fig. 7. Cumulative species curves generated by non-parametric estimators Mao Tau (S ), Chao2,
obi

and Bootstrap for the Odonata larval assemblage from Rio Pinolapa.
298 J.A. Gomez-Anaya & R. Novelo-Gutierrez

winter. This could be

due to the
emergence

of the
pattern sea-

sonal species. Most

of them emerge dur-

ing the dry season (at

the end of the win-

ter and through the

spring). Thus, when

the rains come

in the

few
summer, very spe-
Fig. 8. Cumulative species curves generated by the Clench’s function
cies are present as
lar-
(a = 3.68, b = 0.12, asymptote =
30.67, R =
0.98), and linear depend-
R vae. Summer is the
ence function (a = 2.63, h = 0.102, asymptote = 25.74, = 0.95), for

the Odonata larval assemblage from Rio Pinolapa. The slope was esti- season with a great
2
mated by the first derivative of Clench’s function [a/(l+b*n) ] and the
reproductive activity,
first derivative of Linear dependence function [a*exp(-b*n)].
so that when autumn

comes, there are large populations of larvae of the majority of odonate species.

Later, as time goes on, abundance decreases through mortality until the lower

abundances in spring and then, with emergence of the adults, even lower in sum-

mer. Among strata, the highest species richness was found at shoreline and the

lowest in eddies; while the highest abundance occurred in riffles and the lowest

in eddies. Eddies were

the most simplified stratum, usually with a muddy bottom

and some decaying leaves, and lacking any kind of aquatic plant; this could result

in the lowest richness and abundance. Richness distribution pattern was differ-

ent to that of abundance in RR Thus, 55.39% of all abundance was recorded in

riffles, 34.20% in shorelines and 10.40% in eddies. The most abundant species in

all three strata was Erpetogomphus elaps, with 53.79%, 32.44%, and 13.77%, re-

spectively. This species is, apparently, the best adapted to different conditions in

time and in RP, despite having a

restricted period as
adults (late summer
space

to early autumn).

RENYI’S DIVERSITY PROFILES

Diversity ordering can be performed by using a diversity index family (TOTH-

MERESZ, 1995), as proposed by SOUTHWOOD & HENDERSON (2000). The

Renyi’s diversity profiles method has been poorly used in comparing and order-

ing diversity of aquatic macroinvertebrate assemblages (SIPKAY et al. 2007) and

even less used with Odonata assemblages (JAKAB et al., 2002). In this method a

scale parameter (alpha) is related to the abundance-dominance structure of the

community. At different values of the scale the function is sensitive to rare, com-

and intermediate-abundancespecies. When comparing two Renyi diversity

mon,

profiles, if they do not cross each other, it means that the upper profile represents
 Odonata larval assemblage from Rio Pinolapa 299

a more
diverse assemblage under any common measure of diversity; but if these

profiles do cross each other once, it means that one assemblage is more diverse

when rare species are weighted more heavily (low alpha), and the other assem-

blage is more diverse when common species are weighted more heavily (high al-

pha). When the diversity profiles cross

each other, the communities cannot be

ordered according to their diversity, because one

of them is more
diverse for rare

species, the other for common ones. Renyi’s diversity is sensitive to rare species

for small values of the scale, whereas it is sensitive to abundant species for larger

values of the scale. Seasonal diversity profiles crossed once, mainly for a< 1 val-

ues (Fig. 4). It means that these assemblages are mainly different in rare species.

Strata profiles do not cross; however, shore and riffles profiles were similar in

abundantand frequent species (ct>2), and, definitively, eddies profile was

the least

diverse.

Species distribution must meet the ecological requirements of all stages in the

life cycle (CORBET, 1999). Although the imagoes have more mobility than imma-

ture, adequate substrates must be present for the larval emergence (rocks, vegeta-

tion, twigs, etc.). In the disturbance of the original conditions (e.g.

consequence,

riparian vegetation) will affect richness (SMITH et al., 2007). In RP the major

diversity in shores can be explained because of major microhabitat heterogene-

ity. In shores we observed different size of rocks, sand, algae, plants, roots, mud,

leaf packs, detritus, and different combinations of these substrates. Number of

substrates in riffles and eddies or

combinations of them were
limited. In shores

we found 10 Zygoptera species (1 calopterygid, 1 platystictid, 1 protoneurid, 7

coenagrionids), and 15 Anisoptera (7 gomphids and 8 libellulids). All

Zygoptera

species are endophitic in oviposition and they can find the needed substrates in

shores. Gomphids and libellulids females lack ovipositor and they release their

on
the water surface near
the shores.
eggs

Most of the riffle samples of RP come

from shallow and smooth flowing wa-

ter places. In riffles we found 14 species (5 Zygoptera and 9 Anisoptera). Brech-

morhoga praecox was the most abundant species in this stratum, as have been re-

ported to different species of this genus (CORBET, 1999), and particularly for

this species (BOND et al., 2006). Most of B. praecox larvae were caught, main-

at the end of exhibit different color patterns,

ly, autumn. They disruptive being

cryptic at sand and fine gravel bottom. Must of Coenagrionidae species were re-

corded from shores, except Argia and A. which abundant

oenea tezpi were more

in riffles.

CUMULATIVE CURVES OF SPECIES

The non-parametric estimationachieved by Bootstrap agrees well with the par-

ametric estimation made by Clench function, which predicted 31 species (3 spe-

cies more to
be added). For parametric estimations, we
chose initially the model
300 J.A. Gomez-Anaya & R. Novelo-Gutierrez

2
that better fitted data using the coefficient of determination (R ). This model

was the Clench’s function. Then, we used the slope of the curve evaluated at the

maximum sampling effort by the first derived of both functions, as an approach

of completeness of the list, according to JIMENEZ-VALVERDE & MORTAL

the smaller with the

(2003), and MORTAL & LOBO (2005). Although slope was

Linear dependence believe this function underestimates the number

(0.04), we

of species (since it predicted fewer species than the recorded ones). We think few

species could exist yet to be added to the list for the studied section of the river.

The slope of the first derived of Clench (0.10) is relatively small, supporting the

idea that the list is practically complete.

CCA

Several authors have explored the relationship between some environmental

factors and odonate faunas using multivariate techniques (SAMWAYS, 2003;

SCHINDLER et al„ 2003; HOFFMANN & MASON. 2005; OPPEL. 2005;

CARCHIN1, 2007;

FLENNER & SAH-

SATO &
LEN, 2008;
RIDDIFORD, 2008;

HAMASAKI et al.,

2009). In this work the

CCA showed signifi-

cant correlation be-

tween both environ-

mental and species

matrices. Additional-

ly, the CCA biplot al-

lowed visualizationof

some particular facts

and the establish-

of some
ment hypoth-

eses on species-species
and environment-spe-
cies relationships. For

example, Palaemnema
domina and Argia oe-

nea are
well associat-

Fig. 9. Biplot of CCA ordination showing environmental variables (ar- ed with riffles stratum,

rows) most strongly correlated with axes CC1 and CC2 and species in
indicating similar eco-
triangles. Seasons and strata were included in CCA as dummy variables.

In terms of larval assemblage composition, importantenvi-

logical requirements.
predicting
than less
P. domina is usually
ronmental variables have longer arrows important ones.
 Odonata larval assemblage from Rio Pinolapa 301

found under mid-sized rocks (8-10 cm diam.) in riffles, as

is A. oenea, although
this last species is found also commonly
among rough gravel. Erpetogomphus elaps
is close to the origin on Figure 9, meaning it is the most ubiquitous species in RP.

Apparently, the larvae are well adapted to both erosional and depositional envi-

ronments. Moreover, it was

found in
every
collection and through all seasons.

FINAL CONSIDERATIONS

Contrary to species lists from lentic water bodies (lakes, lagoons, ponds) which

can be considered complete lists because one can sample the whole water body,
the 28 recorded species in the RP survey represent just a part of a bigger assem-

blage. This bigger assemblage extends up and down stream changing in richness,

composition, and abundance with changes in river conditions. From this point
of view, the assemblage here described represents a relatively local measure of

this group of insects, and also a measure of the conservation status of the river.

It is possible that for many species with low abundance and distributionin space

(strata) and time (seasons), the best conditions for their reproduction could be

up or down stream. In this bigger assemblage the floristic, climatic, altitudinal

and microhabitat changes should be considered, because the more

the changes
in the river conditions, the larger the species list should be.

ACKNOWLEDGEMENT

We thank Dr DENNIS R. PAULSON (Seattle) for his criticism and invaluable comments, as well

as the corrections to English syntax.

REFERENCES

ALONSO-EGUlALIS, P.E., 2004. Ecologia de las asociaciones de Odonaia en la zona de influencia

de las microcuencas afectadas por la Zimapdn, Queretaro e Hidalgo. PhD Thesis, Univ.
presa

Auton. Queretaro.

BASELGA, A. & F. NOVOA, 2008. Coleoptera in relict forest of

a
Spain; implications of hyperdi-

for conservation strategies. Ann. enl. Soc. Am. 402-410.

verse taxa 101(2):

BOND, J.G., R. NOVELO-GUTIERREZ, A. ULLOA, J.C. ROJAS, H. QUIROZ-MARTINEZ &

T. WILLIAMS, 2006. Diversity, abundance, and disturbance of Odonata associ-

response

ated with breeding sites of Anopheles pseudopunctipennis (Diptera; Culicidae) in Southern

Mexico. Environ. Enl. 35(6): 1561-1568.

CARCHINI, G, V. DELLA BELLA, A.G. SOLIMINI & M. BAZZANTI, 2007. Relationships be-
tween the presence of odonate species and environmental characteristics in lowland ponds

of central Italy. Ann. Limnol. 43(2): 81-87.

DE CARVALHO, M R., F.A. BOCKMANN, D.S. AMORIM, M. DE VIVO, M. DE TOLEDO-

-PIZA, N.A. MENEZES, J.L. DE FIGUEIREDO, R.M.C. CASTRO, A.C. GILL, J.D.

McEACHRAN, 2005. Revisiting the taxonomic impediment. Science 307: 353.

FLENNER, I. & G. SAHLEN, 2008. Dragonfly community re-organisation in boreal forest lakes:

rapid species turnover driven by climate

change? Insect Consent Divers. 1: 169-179.
 J.A. Gomez-Anaya & R. Novelo-Gutierrez
302

GOMEZ-ANAYA, J.A., R. NOVELO-GUTIERREZ & R. ARCE-PEREZ. 2000. Odonata de la

zonade influencia de la Central Hidroelectrica “Ing. Fernando Hiriart Balderrama” (PH Zi-

mapan), Hidalgo, Mexico, Folia ent. mex. 108: 1-34.

K„ T. K. Y. N. IWASAKI & D.S.

HAMASAKI, YAMANAKA, TANAKA, NAKATANI,

SPRAGUE, 2009. Relative importance of within-habitat environment, land use and

spa-

tial autocorrelations for determining odonate assemblages in rural reservoir ponds in Japan.

Ecol. Res. 24: 597-605.

HAMMOND, P.M., 1994. Practical approaches to the estimation of the extent of biodiversity in

Trans. R. Soc.
speciose groups. Phil. Land. (B) 345: 119-136.

HELLAWELL. J.M., 1986. Biological indicators of freshwaterpollution and environmental manage-

ment. Elsevier, London.

HOFFMANN, T.A. & C.F. MASON, 2005. Habitat characteristics and the distribution of Odonata

in a lowland river catchment in eastern England. Hydrobiologia 539: 137-147.

MORTAL, J. & J. LOBO, 2005. An ED-based protocol for optimal sampling of biodiversity. Biodiv-

ers. Conserv. 14:2913-2947.

JAK.AB, T„ Z. MULLER, G. DEVAI & B. TOTHMERESZ, 2002. Dragonfly assemblages of a

shallow lake reservoir (Tisza-T6, Hungary) and its surroundings. Acta zool. Acad. Sci.
type

hung. 48(3): 161-171.

A. & J. 2003. Las de acumulacion de

JIMENEZ-VALVERDE, MORTAL, curvas especies y la necesi-

dad de evaluar la calidad de los inventarios Rev. Iberica Aracnol. 8: 151-161.

biologicos.
MAGURRAN, A.E., 2004, Measuring biological diversity. Blackwell, Oxford.

MAYR, E„ 1992. A local flora and the biological species concept. Am. J. Bot. 79: 222-238.

McCUNE, B. & XB. GRACE, 2002. Distance Measures. In: B. McCune, XB. Grace & D.L. Ur-

ban, [Eds], Analysis of ecological communities, 45-57. MJM Software Design, Gleneden
pp.

Beach.

2001. Melodos medir la biodiversidad. M&T-Manuales & Tesis Vol. I.

MORENO, C.E., para SEA,

Zaragoza.

MORENO, C.E. & G. HALFFTER, 2000. Assessing the completeness of bat biodiversity invento-

ries usingspecies accumulation curves.

J. appi Ecol. 37: 149-158.

1990. Indicators for hierarchical Conserv. Biol.

MOSS, R.F., monitoring biodiversity: a
approach.

4: 355-364.

NOVELO-GUTIERREZ, R. & E. GONZAlEZ-SORIANO, 1991. Odonata de la Reserva de la Bi-

osfera “La Michilia”, Durango, Mexico, 2: larvas. Folia ent. mex. 81; 107-164.

NOVELO-GUTIERREZ, R„ J.A. GOMEZ-ANAYA AND R. ARCE-PEREZ, 2002. Community

structure of Odonata larvae in two streams in Zimapan, Hidalgo, Mexico, Odonatologica

31(3): 273-286.

OPPEL, S., 2005. Habitat associations of an odonate community in a lower montane rainforest in

Papua New Guinea. Int. J. Odonatol. 2: 243-257.

RESH, V.H., M.J. MYERS & M.J. HANNAFORD, 1996. Macroinvertebrates as biotic indicators

of environmental quality. In: F.R. Hauer & G.A. Lamberti, [Eds], Methods in stream ecology,

pp. 647-667, Acad. Press, San Diego.

SAMWAYS, M J., 2003. Threats to the tropical island dragonfly fauna (Odonata) of Mayotte, Co-

moro archipelago. Biodivers. Conserv 12: 1785-1792.

SATO, M. & N. R1DDIFORD, 2008. A of the Odonata of S’Albufera Natural

preliminary study
Park, Mallorca: status, conservation priorities and bio-indicator potential. J. Insect Conserv

12: 539-548.

M„ C. FESL & A. 2003. Dragonfly associations

SCHINDLER, CHOVANEC, (Insecta, Odonata)
in relation to habitat variables, a multivariate approach. Hydrobiologia 497:169-180.

S1PKAY, C.S., L. HUFNAGEL, A. REVESZ & G. PETRAnYI, 2007. Seasonal dynamics of an

aquatic macroinvertebrate assembly (Hydrobiological case study of Lake Balaton, No. 2).
 Odonata larval assemblage from Rio Pinolapa 303

Appl. Ecol. environ. Res. 5(2): 63-78.

SMITH. J.. M.J. SAMWAYS & S. TAYLOR, 2007. Assessing riparian quality using two complemen-

sets of bioindicators. Biodivers. Conserv. 16: 2695-2713.

tary
SOUTHWOOD. T.R.E. & P.A. HENDERSEN, 2000. Ecological methods. [3rd edn] Blackwell, Ox-

ford.

SPELLERBERG. EE, 1991. Univ. Press.

Monitoring ecological change. Cambridge Cambridge/
UK.

STORK. N.E.. 1994. Inventories of biodiversity; than question of numbers, In: PL.
more a
Forey,
C.J. Humphries & R.l. Vane-Wright, [Eds], Syslemalics and conservation evaluation, SI-
pp.

100. Clarendon Press, Oxford.

TER BRAAK, C.J.F.. 1988. Partial canonical correspondence analysis. In: H.H. Bock, [Ed.], Classi-

fication and related methods of data analysis. pp. 551-558, North-Holland. Amsterdam.

TER BRAAK.C.J.F. & P. SMILAUER. 1998. CANOCOreference manual and user's guide to Canoeo

for Windows: software for canonical community ordination, version 4. Ithasca. NY.

TOTHMERESZ. B,. 1995 Comparison of different methods for diversity ordering. J. Sei. 6:
Veget.
283-290.

TOTHMERESZ, B., 1998. On the characterization of scale-dependent diversity. Ahstr But. 22:

149-156.

WHEELER, Q.D., 2004. Taxonomic triage and the poverty of phylogeny. Philos. Trans. R. Soc.

Land. (B) 359: 571-583.

WHEELER. Q.D., P.H. RAVEN & E.O. WILSON. 2004. Taxonomy: impediment or expedient?

Science 303: 285.