Aquat Ecol DOI 10.1007/s10452-008-9178-7
Intraspecific competition of Glyptotendipes paripes (Diptera: Chironomidae) larvae under laboratory conditions Jan Frouz Æ Richard J. Lobinske Æ Arshad Ali
Received: 24 September 2007 / Accepted: 18 March 2008 Ó Springer Science+Business Media B.V. 2008
Abstract Glyptotendipes paripes larvae were reared in wells of tissue culture plates, in groups of 2, 4, 8, 16, and 32 (representing densities of about 1,300, 2,600, 5,200, 10,400, and 20,800 larvae per m2, respectively). Larval groups were supplied with one of two concentrations (low or high) of food and larvae were individually observed to evaluate the effects of density on mortality, growth, development, behavior, and adult body size. Increased larval densities resulted in higher mortality, as well as slower larval growth and development. The distribution of developmental time became flatter at higher density, with a wider range of values, or even became bimodal. This was a consequence of the most rapidly developing individuals at higher densities emerging as adults sooner than the fastest developing individuals at lower densities, although overall mean developmental time was longer at higher densities. At
J. Frouz A. Ali Department of Entomology and Nematology, Mid Florida Research and Education Center, University of Florida, IFAS, 2725 Binion Rd., Apopka, FL 32703-8504, USA J. Frouz (&) Institute of Soil Biology, Biological Centre, ASCR, Na Sadkach 7, Ceske Budejovice 37005, Czech Republic e-mail:
[email protected] R. J. Lobinske Leon County Mosquito Control and Stormwater Maintenance, 501 Appleyard Dr., Suite A, Tallahassee, FL 32304, USA e-mail:
[email protected]
higher densities, growth and development of smaller larvae were slowed, based on the relative difference in body length between competitors. When larger competitors emerged as adults or died, the growth of smaller larvae may have accelerated, resulting in increased variability of developmental times. The effect of larval density on adult body size was complex, with the largest body size found at the lowest density and a second peak of adult size at high-middle densities, with smaller adult body sizes found at low-middle, and high densities. Similarly, as with developmental time, the range of body size increased with increasing density. Examined food concentrations had no effect on larval mortality, but significantly affected developmental time, growth rate, and adult body size. At higher densities, larvae spent more time gathering food and were engaged in aggressive or antagonistic behaviors. Keywords Behavior Competition Development Food availability Glyptotendipes paripes Larval density
Introduction Intraspecific competition is an important factor that can affect population dynamics and species evolution. Price (1984) described two types of intraspecific competition: equal and scramble. In equal competition, resources become equally divided among competitors and as the numbers of individuals increase, each obtains less of the
123
Aquat Ecol
resources. In scramble competition, resources are not divided equally, and losing competitors die, while winning competitors take the remaining resources, with the greater proportion enhancing their chance for survival. Equal competition is likely to result in unstable population dynamics, with periods of low population densities alternating with massive outbreaks. Scramble competition can result in monopolization of resources by survivors, thus producing a more stable population dynamics. Equal and scramble competition, as described above, form the opposing endpoints of a wide continuum. In reality, competition patterns are somewhere between ‘‘pure’’ equal or ‘‘pure’’ scramble forms. A logical question arises as to what factor(s) determine whether individual organisms become winners or losers in scramble competition. We can intuitively expect that some resources will be shared more or less equally, while others become divided in the scramble pattern. In particular, among some aquatic insects, food filtered from the water column tends to be shared equally by competitors, while use of physical space, such as for the larval tube building exhibited by many chironomid midge species, would be allocated in a win or lose, scramble fashion. In this study, we tested various larval densities at two food concentrations to separate the effects of decreasing space and per capita food availability on competition between laboratory-reared midge larvae. Thus far, relatively more research has been conducted on the effects of interspecific competition in chironomids (e.g., Schmid 1992; Tokeshi 1995), than on intraspecific competition (e.g., McLachlan and McLachlan 1976; McLachlan 1983; McLachlan and Yonow 1989). Adult chironomid midges that emerge in large numbers from urban and suburban lakes in central Florida, USA, may cause severe nuisance, economic losses, and in some situations, medical problems (Ali 1995, 1996). A mid-sized city, Sanford, along the southern shore of Lake Monroe in central Florida has reported economic losses amounting to US 3–4 million dollars annually by waterfront residents, recreationers, visitors, and businesses (Anonymous 1977). Pestiferous midges have been the subject of intensive research efforts in central Florida for more than two decades (Ali 1995, 1996). More recently, the focus of this research has been on the ecology and development of midge larvae (Ali 1990; Ali et al. 2002; Frouz et al. 2002; Lobinske et al. 2002). There are only little data about intraspecific competition among pestiferous
123
midge species (Biever 1971). A better understanding of intraspecific competition mechanisms may be important to develop a clearer picture of competition in general and may also have practical importance. For example, enhanced understanding of the effects of competition on chironomid population dynamics may be incorporated into tools, such as the mathematical model of midge populations from Lobinske et al. (2004), to further improve our understanding or predictive capability. In this study, midges were laboratory-reared in small groups of 2–32 using a technique derived from individual midge rearing designs of Lobinske et al. (2002) and Frouz et al. (2002). This technique enabled observation of the development of individual larva, allowing us to ask questions, such as how developmental rate corresponded with developmental success at the population level. The basic intent of this study was to examine how increasing larval density and different concentrations of food availability affected the developmental rate, growth, survival, and adult size of the pestiferous midge, Glyptotendipes paripes Edwards.
Materials and methods Material collection and preparation Adult G. paripes females were collected in June, 2006 with a sweep net near Lake Harris, Lake County, Florida. In the laboratory, females were released into a 30 9 30 9 30 cm screen cage and allowed to lay eggs overnight in a 15-cm diameter dish that contained 400-ml preaerated 1X Martin’s rearing solution (Martin et al. 1980), supplemented with 1.2 mg l-1 thiamine hydrochloride (Stevens 1998). The next morning, twenty egg masses were collected with a pipette and transferred to a fresh Petri dish containing the rearing solution. These were placed in a growth chamber (Model E-30B, Percival Mfg. Co., Boone, IA) and allowed 48 h to hatch under 30°C constant temperature and white fluorescent light 14:10 h dark photoperiod. Rearing system For midge rearing, clear polystyrene 12-well tissue culture plates (Costar model 3513, Corning Inc.,
Aquat Ecol
Corning, NY) were used. Each well was 22 mm diameter with a flat bottom. Freshly hatched first instar larvae (above) were transferred to individual wells of each plate containing 4 ml of the preaerated rearing solution at larval densities of 2, 4, 8, 16, and 32 per well; these densities represented nearly 1,300, 2,600, 5,200, 10,400, and 20,800 larvae per m2. The bottom of each well was covered by a thin layer of sand, about 1 grain thick. For the experiment, 30 wells were used for each larval density and two food concentrations (high and low) combination tested. Each tissue culture plate contained both concentrations of food. Using a latin square design, one half of the plate was assigned to low concentration and the other half to high food concentration with each concentration having full
range of larval density. However, some replicates were lost by accident (cracked plates) or occasional handling errors (mostly accidental removal of some larvae during liquid replacement). Data from such plates were excluded from further processing and therefore, the actual number of replicates for individual treatments is presented in Table 1. The rearing plates were maintained in a growth chamber under temperature and light conditions described above. The rearing medium in each well was replaced daily with 4 ml of fresh, preaerated, 19 Martin’s solution. After media replacement, a food suspension was added, either 35 ll (low food concentration) or 70 ll (high food concentration) per well. The food suspension was prepared by using
Table 1 Mean ± SEM percentage of introduced Glyptotendipes paripes larvae that successfully completed development to adult emergence and proportions of larvae that survived the
first 5 days of development, reared on two food concentrations (low and high) in the laboratory
Larval density
Proportion of successful larvae %
Five day larval survival %
Number of replicates
Low food concentration 2
33.9 ± 6.8a
60.7 ± 7.9
28
4
14.1 ± 3.5b
56.5 ± 6.5
23
8
8.9 ± 1.8b
60.9 ± 5.6
24
16
2.7 ± 0.9b
56.9 ± 4.6
28
32
2.5 ± 0.9b
59.9 ± 6.0
13
High food concentration 2
16.7 ± 5.3a
48.1 ± 8.2
27
4
19.6 ± 4.1a
63.0 ± 5.4
23
8
10.3 ± 1.7ab
58.2 ± 5.2
23
16 32
3.6 ± 1.1b 2.3 ± 1.1b
40.1 ± 4.9 46.5 ± 5.2
26 16
Both food concentrations pooled 2
25.5 ± 4.5a
54.5 ± 5.7
55
4
16.8 ± 2.7ab
59.8 ± 4.2
46
8
9.6 ± 1.2b
59.6 ± 3.8
47
16
3.1 ± 0.7c
48.8 ± 3.5
54
32
2.4 ± 0.7c
52.5 ± 4.0
29
Two-way ANOVA
Proportion of successful larvae %
Five day larval survival %
df
F
P
F
P
Larval density
4
8.55
Food conc.
1
0.53
\0.0001
1.18
0.3182
0.4635
3.57
Interaction
4
2.66
0.0599
0.0335
1.15
0.3324
F and P values of two-way ANOVA for each parameter are shown at the bottom of the table; df values are the same for successful larval development and for first 5 day larval survival. Mean values of percent developmental success for various larval densities in a column under low, high, and pooled food concentrations followed by the same letter are not significantly different (one-way ANOVA, LSD test, P \ 0.05)
123
Aquat Ecol
100 mg of Fish Starter Diet (Aquatic Ecosystems, Apopka, Florida, USA) and 20 mg of powdered dry baker’s yeast in 4 ml of Martin’s rearing solution. Typically 20–28 ml of the food suspension was made and maintained with a magnetic stirrer while aliquots of required volume were taken with a suitable micropipette for each feeding. Developmental and behavioral observations During daily media replacement, the wells were examined and any midge developmental changes or mortality were recorded. Cadavers of dead larvae were removed at this time. Larval molting to the next instar was detected based on the presence of associated exuviae, as well as through observing the larval head as wide or wider than the thorax at the start of each instar, and the head being much narrower than the thorax at the end of each instar. These changes were assumed to occur at the mid-point between the two successive observations. Individual larvae were identified according to their occupancy of a particular tube under the assumption that a larva spends all its developmental time in one tube. To distinguish individual larvae, a map of larval tubes and their occupancy was constructed for each well and updated daily. In addition to these daily observations, the length of larvae was measured weekly for the first 5 weeks of the experiment using an eyepiece-mounted micrometer fitted to a stereomicroscope (Baush & Lomb, USA; 100 lm accuracy) using light from underneath to make larvae in the tubes easily visible. Upon adult emergence, sex was determined and each adult preserved in 70% ethanol. The left wing of each adult was individually mounted on a temporary slide and wing length from humerus to apex was measured with an eyepiecemounted micrometer fitted to a compound microscope (Model BHA, Olympus Corp., Tokyo, Japan, 100 lm accuracy). The experiment was terminated after 70 days at which time all larvae had either emerged as adults or died. Seven days after hatching, the number of larvae that formed tubes and the number of larvae that were freeswimming in the water column, without tubes, were counted. Larvae were distinguished as swimming based on characteristic behavior consisting of episodes of the larva arching its back into an omega-shaped loop and then suddenly extending the body straight. Behavioral observations of larval feeding were made at 21 days after hatching, with at least 10 randomly
123
selected larvae from 10 wells observed for two minutes for each larval density and food concentration combination. The proportion of time each larva spent crawling, undulating, staying calm, or trying to cannibalize other larvae was recorded. A ‘crawling’ larva extended its body partly or completely from its tube and usually scratched for food from the bottom of the well. An ‘undulating’ larva performed wave-like movements with its body inside the tube, while ‘calm’ meant that the larva stayed in the tube and did not perform any visible movement; cannibalizing meant that the larva was attempting to bite other larvae and/or trying to push them out of their tubes. Statistical analyses A two-way analysis of variance was used to compare the effects of larval density and food concentration on larval survival 5 days after emergence and the percentage of larvae successfully completing development. For this purpose, all specimens that emerged from pupae were counted as successful irrespective of their wing expansion. The limited space between the plate lid and liquid surface in the wells sometimes interfered with adult emergence. Specimens that did not develop to adults were classified as unsuccessful. The percentage of successful larvae was thus the inverse of total mortality (100-mortality). A square root arcsine transformation was used on percentage data values, such as larval survival or successful development, prior to subjecting the data to ANOVA. For other parameters, individual larvae were considered as a replicate. Seventy days after the start of the experiment, all introduced larvae had emerged as adults or died. By following the records of individual larva backwards, we could distinguish successful larvae from those that were not, and among successful larvae, distinguish males and females. A three-way ANOVA was applied to analyze the effects of additional parameters, such as sex, and their combinations. An LSD post-hoc test was used after ANOVA to distinguish treatment effects in significant factors. General linear models were applied to explore factors affecting larval behavior in reference to some continual predictors (such as, actual number of larvae per well). Repeated measures ANOVA was used to compare larval growth at different densities and food levels. Growth rate was defined as an increase in larval size over time.
Aquat Ecol
Spearman’s rank correlation was used to explore the correlations between actual number of larvae per well and the proportion of time dedicated to certain behaviors. SPSS 10.0 was used for statistical analyses.
Results The percentage of larvae that successfully completed development decreased with increasing larval density
(Table 1). There was a significant interaction between larval density and food concentration; at low food concentration there was a rapid drop in developmental success between densities of two and four larvae per well. At high food concentration, developmental success did not differ significantly between two and four larvae per well, but decreased sharply at higher larval densities. The highest mortality occurred at the start of experiment (Fig. 1). In all treatments, about half of the larvae died within 5 days after hatching.
Fig. 1 Larval survival (line) and adult emergence at 5-day intervals (bars) of laboratory-reared Glyptotendipes paripes at five different larval densities and two food concentrations: (a) low food concentration, 2 larvae per well; (b) high food concentration, 2 larvae per well; (c) low food concentration, 4 larvae per well; (d) high food concentration, 4 larvae per well; (e) low food concentration, 8 larvae per well; (f) high food concentration, 8 larvae per well; (g) low food concentration, 16 larvae per well; (h) high food concentration, 16 larvae per well; (i) low food concentration, 32 larvae per well; (j) high food concentration, 32 larvae per well. Mean ± SEM of adult emergence (days) and kurtosis for emergence distribution is given
123
Aquat Ecol
density but different concentration of food supply were compared, we found that food supply had a significant effect (GLM, F = 11.09, P = 0.0011, n = 147, larval density used as a co-variable, both sexes pooled). With increasing larval density, the distribution of developmental time became flatter (platykurtic), as documented by lower kurtosis values. These values actually became negative for larval densities of 16 and 32 larvae per well. At these high densities, the developmental time distribution became bimodal in some cases. This bimodality corresponds with the uneven development of larvae in individual wells with higher larval densities, when some larvae developed faster than others; in slower developing larvae the instar time extended until the faster growing larvae either emerged as adults or died (Fig. 2). This resulted in a wide range of developmental times (Table 2). Larvae that successfully completed development grew faster during the second instar than those that were not successful in terms of development (Table 3). Larval
The 5 day mortality percentage was independent of density or food concentration (Table 1). Developmental time significantly increased with increasing larval density and decreased under high food concentration (Table 2). Males developed significantly faster than females. Clearly, density and food supply were correlated, because at higher densities, larvae had lower per capita food supply. However, because the high food supply rate was double that of the low food supply and the experimental larval densities increased 29 per step, this created several pairs of experimental larval groups with the same per capita food supply. For example, two larvae per well with low food supply and four larvae per well with high food supply had the same per capita food availability. When these matching per capita food supply pairs were examined, we found no significant effect of density (GLM, F = 1.01, P = 0.3149, n = 126, per capita food level used as a co-variable, both sexes pooled). In contrast, when pairs of larval groups with the same
Table 2 Mean ± SEM and range of duration of developmental time (days) of Glyptotendipes paripes reared in the laboratory at two food concentrations and five larval densities Parameter
Number of larvae per well 2
4
8
16
32
54.5 ± 3.7(4)
Low food concentration Male Female
41.2 ± 3.0 (8)
41.3 ± 3.3(8)
39.1 ± 3.3(8)
47.6 ± 5.5(9)
34–62
27–59
27–57
31–78
42–61
42.4 ± 1.4(9)
45.5 ± 2.0(6)
45.1 ± 3.2(9)
51.3 ± 7.3(3)
53.8 ± 3.8(8)
35–47
37–51
36–68
35–66
43–75 48.6 ± 3.4(7)c
High food concentration Male Female
30.1 ± 0.8(7)a
38.1 ± 3.5(9)b
36.1 ± 1.8(9)ab
38.6 ± 1.8(8)b
26–33
24–62
29–45
30–46
42–66
34.0 ± 1.4(2)a
38.0 ± 1.4(10)a
41.6 ± 3.0(11)ab
47.3 ± 2.3(7)bc
54.8 ± 3.1(4)c
32–36
32–47
29–63
37–57
44–59
Both sex and food concentrations pooled 36.9 ± 1.5(24)a
40.7 ± 1.5(33)a
40.7 ± 1.6(37)a
46.2 ± 2.3(29)b
52.7 ± 1.9(23)c
26–62
24–62
27–63
30–78
42–75
Three-way ANOVA
n
df
F
P
Larval density
147
4
9.46
\0.0001
Food conc.
147
1
10.10
0.0018
Sex
147
1
4.81
0.0300
Numbers in parentheses are the actual number of replicates used in calculations; full factorial design was performed and non-significant interactions omitted from the table. Mean values in a row with the same letter following are not significantly different (one-way ANOVA, LSD test, P \ 0.05; values without any letter were not significantly different)
123
Aquat Ecol
Fig. 2 Diagrammatic presentation of Glyptotendipes paripes larval positions, larval size, and development in one experimental well (replicate) with 16 introduced larvae and low food concentration. Numbers below circles represent days since hatching. Lines
represent tube positions, abandoned tubes are marked by dashed lines. Arabic numerals represent larval length in mm while Roman numerals represent larval instar. Number accompanying male or female symbol is duration of development (days)
density significantly affected larval growth (Fig. 3); post-hoc LSD comparison shows that 32 per well grew significantly slower than those reared at other densities. No effect of food or food density interaction was detected by repeated measures ANOVA (Fig. 3). Table 4 summarizes the comparison of body length of 28-day-old larvae grown at various densities. Larval size is significantly affected by density and developmental success, with successful larvae, and larvae reared at lower densities, growing faster. There was significant interaction between density and successful development. At lower density, there was no significant difference between the sizes of successful and unsuccessful larvae; however, with increasing density the difference between successful and unsuccessful larvae increased. Size of unsuccessful larvae decreased with increasing density (Table 4), while the size of successful larvae did not decrease at densities lower than 16 or 32 per well. At high food concentration, larvae reared at higher densities were even larger than larvae reared at lower density (Table 4). At the higher densities, larval size distribution became flatter and only larger larvae were successful (Fig. 4). Adults with the longest mean wing length emerged from the two larvae per well density. Mean wing length significantly decreased for adults that emerged from four larvae per well density, but increased again for adults emerging from 8 and 16 larvae per well densities (Table 5).
The construction of tubes by young larvae was strongly affected by larval density. At densities of 2 and 4 larvae per well, most larvae built tubes within 1 week, while almost half of the larvae reared at 32 larvae per well continued to move freely in the water without building tubes (Table 6). No significant effect of food supply on the larval tube building behavior was detected. Larval density also significantly affected behavior of the investigated larvae (Table 7). At day 21 of development, the proportion of undulating time of larvae in the tubes significantly increased with increasing larval density, while the proportion of larvae that stayed calm decreased (Table 7, Fig. 5). The proportion of time dedicated by larvae to crawling behavior was significantly affected by the number of larvae in the well and peaked at 8 larvae per well. In wells with more larvae, the incidence of cannibalistic behavior also increased (Spearman’s rank correlation between number of larvae and attack occurrence per well, 0.450, P \ 0.05, n = 49).
Discussion Larval development at two larvae per well and high food concentration was comparable to the development of G. paripes larvae reared individually under similar laboratory conditions and the same per capita
123
Aquat Ecol Table 3 Mean ± SEM and range of time (days) to complete development to adult emergence for individual larval instars of Glyptotendipes paripes reared in the laboratory at two food concentrations and five larval densities Larval density
1st Instar
2nd Instar
Mean ± SEM(n)
Range
3rd Instar
Mean ± SEM(n)
Range
4th Instar
Mean ± SEM(n)
Range
Mean ± SEM(n)
Range
Successful larvae (low food concentration) 2
2.3 ± 0.(17)
1–3
6.4 ± 0.6(17)
3–12
8.1 ± 0.7(17)
4–15
23.3 ± 1.6(17)
13–41
4
2.3 ± 0.1(14)
2–3
6.4 ± 0.7(14)
4–13
8.6 ± 0.7(14)
5–13
24.8 ± 2.4(14)
10–41
8
2.3 ± 0.1(17)
2–3
6.9 ± 0.6(17)
3–13
7.6 ± 0.8(17)
3–16
24.3 ± 2.4(17)
8–53
16
2.7 ± 0.2(12)
2–4
7.2 ± 0.7(12)
3–12
10.6 ± 1.4(12)
5–22
26 ± 4.1(12)
10–58
32
3.1 ± 0.1(12)
3–4
8.5 ± 0.6(12)
5–13
13.7 ± 1.5(12)
4–20
27.3 ± 2.2(12)
11–39
Successful larvae (high food concentration) 2
2.0 ± 0.3(9)
1–3
7.4 ± 0.4(9)
6–9
7.0 ± 0.9(9)
4–14
13.5 ± 0.9(9)
9–18
4
2.6 ± 0.1(19)
2–3
7.4 ± 0.4(19)
4–13
7.8 ± 0.5(19)
5–14
19.1 ± 1.6(19)
7–41
8
2.6 ± 0.2(20)
2–6
7.1 ± 0.5(20)
4–13
7.6 ± 0.6(20)
4–13
20.7 ± 1.9(20)
7–47
16 32
2.4 ± 0.1(15) 5.0 ± 0.4(11)
2–3 4–8
6.9 ± 0.5(15) 6.0 ± 0.6(11)
4–12 4–11
9.5 ± 0.8(15) 9.9 ± 1.2(11)
5–14 4–17
22.7 ± 2.0(15) 23.7 ± 3.8(11)
8–38 5–51
Unsuccessful larvae (low food concentration) 2
2.4 ± 0.1(19)
2–3
8.4 ± 0.8(12)
6–14
7.0 ± 0.8(8)
3–12
22.6 ± 4.2(3)
13–31
4
2.4 ± 0.1(45)
2–3
8.1 ± 0.6(32)
3–19
9.0 ± 0.8(22)
5–21
27.3 ± 3.2(6)
17–37
8
2.5 ± 0.1(98)
2–5
9.6 ± 0.5(158)
3–23
8.0 ± 0.7(36)
3–22
29.7 ± 3.9(9)
16–49
16
2.8 ± 0.1(258)
2–6
10.2 ± 0.6(110)
1–43
8.8 ± 0.6(59)
3–25
34.5 ± 4.3(6)
34–52
32
3.4 ± 0.1(267)
2–6
11.2 ± 0.2(95)
5–15
14.8 ± 2.1(26)
5–48
34.0 ± 2.2(6)
25–40
Unsuccessful larvae (high food concentration) 2
2.2 ± 0.1(17)
1–3
7.5 ± 1.0(6)
6–13
7.0 ± 0.4(6)
6–9
17.0 ± 0(1)
17
4
2.7 ± 0.1(34)
2–4
8.2 ± 0.4(30)
5–13
7.7 ± 0.8(23)
5–25
18.1 ± 1.9(8)
10–29
8
2.6 ± 0.1(92)
2–9
9.1 ± 0.4(53)
3–14
8.2 ± 0.7(29)
3–24
26.7 ± 0.9(3)
25–29
16
2.7 ± 0.1(203)
2–6
8.2 ± 0.4(38)
1–13
8.8 ± 1.2(17)
4–24
18.2 ± 0.4(4)
17–19
32
4.9 ± 0.1(81)
4–6
8.6 ± 0.6(26)
4–13
6–20
28.0 ± 2.8(2)
24–32
ANOVA
df
1st Instar
12.3 ± 2.3(6)
2nd Instar
F
P
F
3rd Instar
4th Instar
P
F
P
F
P
Density Food
4 1
18.53 11.77
\0.0001 0.0006
0.62 2.56
0.6430 0.1094
4.79 0.66
0.0008 0.4163
1.68 13.11
0.1569 0.0003
Success
1
2.55
0.1101
14.32
0.0001
0.06
0.8038
3.06
0.0820
Full factorial three-way ANOVA was performed and non-significant interactions omitted from the table. Successful larvae emerged as adults, unsuccessful larvae died before emergence
food density (Lobinske et al. 2002), indicating that at those conditions, intraspecific larval competition was low. In agreement with several developmental reports on insects (e.g., Renshaw et al. 1993; Teng and Apperson 2000; Hooper et al. 2003; Frouz et al. 2004b; Ireland and Turner 2006), we found that G. paripes larval mortality increased with increasing larval density, while larval growth and development slowed. Per capita food density did not influence
123
mortality but led to increased developmental duration. Competition for space in G. paripes appears to fit the scramble hypothesis, whilst competition for food follows the equal competition model. Higher larval density may affect life cycle parameters either through decreased per capita food density or by other density dependent factors, including larval aggression. Higher density also increases larval activity (Fig. 5, Table 5), which makes larval persistence
Aquat Ecol
Fig. 4 Comparison of the body length distributions of successful (emerged as adults) and unsuccessful (died before adult emergence) 28-day-old laboratory-reared Glyptotendipes paripes larvae grown at low food concentrations and at densities of 2 (a) and 16 (b) larvae per well. ***Significantly different by Student’s t-test, P \ 0.001
Fig. 3 Mean ± SEM body length of laboratory-reared Glyptotendipes paripes at five larval densities and two food concentrations: (a) low food concentration; (b) high food concentration
(Bedhomme et al. 2005), reduce available oxygen, and alter chemical conditions, all of which can slow or otherwise affect larval growth.
energetically more costly. Higher larval density may also increase production of various metabolic byproducts which may slow the growth of larvae in the well
Table 4 Mean ± SEM body length (mm) of 28-day-old Glyptotendipes paripes larvae reared in the laboratory at two food concentrations and five larval densities Parameter
Number of larvae per well 2
4
8
16
32
Low food concentration Successful
9.28 ± 0.25a
9.36 ± 0.37a
9.20 ± 0.23a**
8.53 ± 0.17a**
Unsuccessful
9.00 ± 0.22a (10)
8.42 ± 0.29a (19)
7.51 ± 0.23b (35)
6.53 ± 0.25ab (52)
6.8 ± 0.23b** 5.89 ± 0.18b (64)
High food concentration Successful
8.33 ± 0.56b
9.63 ± 0.27a**
9.33 ± 0.17ab**
9.00 ± 0.30ab
7.00 ± 0.28c**
Unsuccessful
9.66 ± 0.54a (3)
8.25 ± 0.33a (16)
7.36 ± 0.32b (22)
8.12 ± 0.32ab (16)
5.66 ± 0.18c (33)
Three-way ANOVA
df
F
P
Density
4
34.5
0.2180
Success
1
44.5
\0.0001 \0.0001
Food
1
1.5
Food and density
4
2.4
0.0478
Density and success
4
5.5
0.0002
Numbers in parentheses represent number of replicates (larvae). ** Successful larvae were significantly larger than unsuccessful larvae (Student’s t-test, P \ 0.01). Statistically homogeneous groups of means in one row are followed by the same letter (one-way ANOVA, LSD test, P \ 0.05). The output of three-way ANOVA, comparing effects of larval density, successful adult emergence, food concentration, and their interactions are summarized at the bottom of the table, full factorial design was performed and non-significant interactions omitted from the table
123
Aquat Ecol Table 5 Mean ± SEM wing length (mm) and range, in parentheses, of male and female Glyptotendipes paripes reared in the laboratory at two food concentrations and five larval densities Parameter
Number of larvae per well 2
4
8
16
32
Low food concentration Male Female
3.55 ± 0.12
3.10 ± 0.16
3.55 ± 0.12
3.24 ± 0.16
3.25 ± 0.15
(3.30–3.75)
(2.60–3.60)
(2.75–3.70)
(3.10–3.60)
(2.90–3.50)
3.32 ± 0.15
3.18 ± 0.16
3.27 ± 0.16
3.35 ± 0.14
3.21 ± 0.18
(3.00–3.65)
(2.90–3.45)
(2.90–3.80)
(3.10–3.70)
(2.60–3.70)
High food concentration Male Female
3.54 ± 0.06
3.49 ± 0.12
3.53 ± 0.12
3.35 ± 0.23
3.25 ± 0.17
(3.25–3.75)
(2.85–4.10)
(3.10–4.00)
(3.00–4.50)
(2.70–3.90)
3.67 ± 0.10
3.31 ± 0.18
3.67 ± 0.09
3.67 ± 0.23
3.33 ± 0.20
(3.50–3.90)
(2.80–3.75)
(3.25–4.20)
(2.80–4.00)
(3.00–3.70)
Both sex and food concentrations pooled 3.48 ± 0.15 (3.00–3.90)
3.28 ± 0.19
3.43 ± 0.18
3.43 ± 0.18
3.23 ± .18
(2.60–4.10)
(2.75–4.20)
(2.80–4.50)
(2.60–3.90)
Three-way ANOVA
n
df
F
P
Larval density Food conc.
147 147
4 1
3.05 0.36
0.0194 0.5486
Sex
147
1
11.81
0.0008
The output of three-way ANOVA are summarized at the bottom of the table, full factorial design was performed and non-significant interactions omitted from the table
Table 6 Mean ± SEM of percentage of free swimming (without tubes), 7-day-old Glyptotendipes paripes larvae reared in the laboratory at two food concentrations and five larval densities; n is the number of wells considered Food concentration
Number of larvae per well 2
4
8
16
32
Low
0.0 ± 0.0a
0.0 ± 0.0a
26.2 ± 12.7b
26.8 ± 6.9b
58.3 ± 3.4c
High
6.3 ± 6.3ab
0.0 ± 0.0a
17.5 ± 8.5b
45.0 ± 5.0c
44.0 ± 4.0c
Both pooled
3.1 ± 3.2a
0.0 ± 0.0a
21.8 ± 7.9b
35.9 ± 5.3c
51.2 ± 3.4d
n
df
F
P
Larval density
70
4
18.05
\0.0001
Food conc.
70
1
1.39
0.2452
Parameter Two-way ANOVA
Means in a row with the same letter following are not statistically different by one-way ANOVA (LSD test, P \ 0.05, full factorial design was performed but non-significant interactions omitted from the table)
In our experiment, five larval densities and two food concentrations were used to separate the effects of per capita food availability and associated influences of larval density. Larval survival was highly dependent on larval density and this parameter seemed to have a greater influence on larval survival
123
than food availability (Table 1). The proportion of swimming larvae that did not build tubes was strongly density dependent, but independent of food availability (Table 6). Swimming, where the larvae can be transported by water movement, is a principal method of dispersion of young chironomid larvae in
Aquat Ecol Table 7 Output of general linear models comparing effect of the actual number of Glyptotendipes paripes larvae in individual wells, original number of larvae introduced to Parameter
Behavior Calm
Actual numbers
individual wells, and food concentration, on the proportion of time dedicated to certain behaviors, non-significant interactions were omitted from the table
Undulating
Crawling
F
P
F
P
F
P
15.6
0.0001
7.0
0.0090
0.21
0.6473
Original numbers
0.6
0.6131
0.8
0.5188
2.86
0.0395
Food concentration n
1.8
0.1451 135
1.6
0.1892 135
0.30
0.8242 135
Spearman’s rank correlation
-0.5228
0.4176
0.0774
‘‘n’’ represents the actual total number of larvae considered for both GLM and correlation. Spearman rank correlation represents the correlation between actual number of larvae per well and the proportion of time dedicated to certain behaviors
Fig. 5 Mean ± SEM of proportion of time for which 21-dayold Glyptotendipes paripes larvae exhibited specific behaviors, such as: (a) calm, no visible movement; (b) undulating movement of body in the tube; (c) crawling, extended from tube or outside tube
the field (Pinder 1995). Therefore, prolongation of swimming activity at high densities can be interpreted as an adaptation against overcrowding, because under natural conditions, swimming larvae would migrate out of an overcrowded area. This shift in larval behavior could be initiated either by
mechanical signals, such as the number of contacts with conspecific larvae, or in response to chemicals released in water (Bedhomme et al. 2005). The number of larvae that a patch of habitat supports can be affected also by sediment quality; Frouz et al. (2004a) showed that sediments that are suitable for burrowing and making vertical tubes may support more larvae than hard sediments where larvae made horizontal tubes on sediment surface. Larval mortality was highest during the initial stages of development just after larval hatch. During this period, about half of all test larvae died, and this mortality seemed to be density independent. This is consistent with other studies indicating that density independent mortality may strongly affect overall mortality and population dynamics (Lozano et al. 1993; Cooke and Roland 2003). The reason for such high mortality in early developmental stages is probably connected with the demanding shift from planktonic to benthic dwelling and associated tube building. Production of tubes is demanding for mature larvae (McKie 2004) and it may be more pronounced in small larvae because of larger surface volume ratios. Tube building represents important energetic spending for young larva which may represent a bottle-neck for larval survival. In contrast to survival rate, the duration of larval development seems to be strongly affected by both larval density and food supply. The principal effects of per capita food availability on developmental duration in our study is in agreement with other reports (e.g., Sankarperumal and Pandian 1991; Teng and Apperson 2000). In agreement with Frouz et al.
123
Aquat Ecol
(2002), males developed faster than females. However, no significant interactions between sex and density, or sex and food supply in developmental time were detected, indicating that response of both sexes to increasing competition did not differ. Looking at individual instars, the developmental duration of the fourth instar, which typically represents the majority of overall developmental time (Frouz et al. 2002; Lobinske et al. 2002), strongly depends on food density. On the other hand, duration of the first instar is strongly density dependent, with longer duration of this instar at higher densities. This may correspond with the longer persistence of swimming behavior at higher densities discussed earlier. Duration of the second instar was significantly different between larvae that successfully finished development to adults and those that failed to develop to adult stage. This indicates that the developmental time duration of the second instar and the time when the larva reaches third instar could be critical for successful adult emergence. Third instar is a time when larval growth accelerates (Fig. 2), so it is very likely that larvae developing to third instar sooner than others will gain an advantage over their competitors. The reason why growth in higher instars did not differ between successful and unsuccessful larvae may be that some slow growing larvae may also complete development (Fig. 2). In agreement with Rasmussen (1985) and Frouz et al. (2004b), the investigated larvae developed more slowly at higher larval densities, and growth rate became more variable at these densities; however, the biggest larvae at higher densities may grow faster than the biggest larvae at lower densities, although growth in the majority of larvae slows down with increasing density. Although many authors have reported decreasing adult body size with increasing larval density (Renshaw et al. 1993; Ireland and Turner 2006), in our study, the effect of larval density on adult body size was more complex. The largest adult size (based on wing length) was recorded at the lowest density, decreased at lowmedium density, increased at high-moderate density, and decreased again at high density. This phenomenon seems to correspond with the broader range and flatter distribution of developmental times and body sizes that increased with density. At higher density, the distribution of developmental times even became bimodal. As a consequence, the individuals that developed most rapidly at higher densities may have done so more quickly than the fastest developing
123
individuals at lower densities, even though mean developmental time was longer at the higher density (Fig. 1). Similarly, the variance of body size increased with larval density, thus typically at higher densities the largest larvae were as long, or longer, than the largest larvae at low densities (Table 3). This may explain the complex pattern of body size changes. At lower density conditions, mortality equally affected all size groups of larvae, while at higher densities, larger larvae were more likely to survive (Fig. 4). This resulted in mean body size remaining high at intermediate densities. McLachlan (1983) reported increases in body size with increasing density as a consequence of high mortality that affected mainly smaller larvae. At very high densities, however, even for the largest specimens, body size decreased, as did overall average body size. An interesting question is: how is it possible that larval body size and adult body size of midges reared at intermediate high density remain the same or even increase in comparison with those reared at low density, even though per capita food supply is much lower for the former group? Certainly, part of the explanation can be prolongation of developmental time. This is insignificant at densities of 4 and 8 larvae per well, but starts to be more pronounced at larval densities of 16 or higher per well (Table 2). McLachlan (1983) explained faster growth of successful competitors by extra food provided by bodies of dead competitors. We cannot completely exclude this food source in our experiment. However, we believe that in this situation, when cadavers were removed daily with water replacement, the role of this food supply is limited. We propose three alternative, non-exclusive hypotheses to explain why some larvae can grow faster at lower food density. The presence of competing larvae can bring benefits that may save energy for larvae, such as successful larvae using the tubes of dead competitors and thus saving some energy on tube production. Another possibility is that at higher densities, larvae somehow switch their metabolism to support more efficient food use and faster growth. We expect that there should be some price in terms of future cost to ‘turn on the metabolic switch,’ otherwise larvae would grow at high rates all the time. Finally, it may be the question of variability and selection in the population. We can expect that ability to grow fast has some dependence on distribution of specific traits within a population. At low density, fast growth may not be so critical and hence is
Aquat Ecol
not subject to strong selection. As the density increases, selection for faster growing individuals increases as well, with a simultaneous increase in number of specimens that are available for selection. The bimodal distribution of larval developmental time at higher densities was caused by variability in developmental rate between larvae. At higher densities, growth and development of smaller larvae slowed down until the time larger competitors emerged as adults, or died; thereafter, growth and development of the smaller larvae resumed. This increased the variability of developmental times and, in some cases, yielded bimodal distributions of adult emergence (Figs. 1 and 2). Such a postponement of development by smaller larvae has been mentioned by Biever (1971). Our data show that at high densities, these postponed larvae can form a substantial part of a population’s overall emergence (Fig. 1g, Fig. 2). Thus, at higher densities, two peaks of emergence may occur from one larval cohort. This may be a reason why the developmental time of G. paripes, based on the daily abundance pattern of adult emergence reported by Ali et al. (1985) was shorter than that found by rearing individual larvae under laboratory conditions (Lobinske et al. 2002). The persistence of some slower growing larvae at the time when first adults of their cohort may be ready to lay eggs provides evidence of overlapping generations. It may also result in complex intraspecific interactions, such as competition between old larvae and young larvae of a new generation that may affect the overall population dynamics. These interactions clearly require more attention in future research. Acknowledgments This study was undertaken while the senior author, J. Frouz, was on a Fulbright fellowship at the University of Florida; sincere gratitude is expressed to the Fulbright Foundation for the research support. The study was partly supported by the Research Plan ISB BC AS CR No Z6066911. Three anonymous reviewers are thanked for their critique which has improved overall quality of the article.
References Ali A (1990) Seasonal changes of larval food and feeding of Chironomus crassicaudatus (Diptera: Chironomidae) in a subtropical lake. J Am Mosq Control Assoc 6:84–88 Ali A (1995) Nuisance, economic impact and possibilities for control. In: Armitage PD, Cranston PS, Pinder LCV (eds) The Chironomidae: the biology and ecology of non-biting midges. Chapman & Hall, London, UK, pp 339–364
Ali A (1996) Pestiferous Chironomidae and their management. In: Rosen D, Bennett FD, Capinera JL (eds) Pest management in the subtropics: integrated pest management—a Florida perspective. Intercept, Andover, UK, pp 487–513 Ali A, Stanley BH, Majori G (1985) Daily abundance patterns of pestiferous Chironomidae (Diptera) in an urban lakefront in central Florida. Environ Entomol 14:780–784 Ali A, Frouz J, Lobinske RJ (2002) Spatio-temporal effects of selected physico-chemical variables of water, algae, and sediment chemistry on larval community of nuisance Chironomidae (Diptera) in a natural and a man-made lake in central Florida. Hydrobiologia 470:181–193 Anonymous (1977) Economic impact statement, blind mosquito (midge) task force. Sanford Chamber of Commerce, Seminole Co., Florida Bedhomme S, Agnew P, Sidobre C, Michalakis Y (2005) Pollution by conspecifics as a component of intraspecific competition among Aedes aegypti larvae. Ecol Entomol 30:1–7 Biever KD (1971) Effect of diet and competition in laboratory rearing of chironomid midges. Ann Entomol Soc Am 64:1166–1169 Cooke BJ, Roland J (2003) The effect of winter temperature on forest tent caterpillar (Lepidoptera: Lasiocampidae) egg survival and population dynamics in northern climates. Environ Entomol 32:299–311 Frouz J, Ali A, Lobinske RJ (2002) Influence of temperature on developmental rate, wing length and larval head capsule size of pestiferous midge Chironomus crassicaudatus (Diptera: Chironomidae). J Econ Entomol 95:699–705 Frouz J, Lobinske RJ, Ali A (2004a) Influence of Chironomidae (Diptera) faecal pellet accumulation on lake sediment quality and larval abundance of pestiferous midge Glyptotendipes paripes. Hydrobiologia 518:169–177 Frouz J, Ali A, Lobinske RJ (2004b) Laboratory evaluation of six algal species for larval nutritional suitability of the pestiferous midge Glyptotendipes paripes (Diptera: Chironomidae). J Econ Entomol 97:1884–1890 Hooper HL, Sibly RM, Maund SJ, Hutchinson TH (2003) The joint effects of larval density and 14C-cypermethrin on the life history and population growth rate of the midge Chironomus riparius. J Appl Ecol 40:1049–1059 Ireland S, Turner B (2006) The effects of larval crowding and food type on the size and development of the blowfly, Calliphora vomitoria. Forensic Sci Int 159:175–181 Lobinske RJ, Ali A, Frouz J (2002) Laboratory estimation of degree-day developmental requirements of Glyptotendipes paripes (Diptera: Chironomidae). Environ Entomol 31:608–611 Lobinske RJ, Stimac JL, Ali A (2004) A spatially explicit computer model for immature distributions of Glyptotendipes paripes (Diptera: Chironomidae) in central Florida lakes. Hydrobiologia 519:19–27 Lozano C, Kid NAC, Campos M (1993) Studies on the population dynamic of bark beetle, Leperinus varius (Fabr.) (Coleoptera: Scolitidae) on European Olive (Olea europea). J Appl Entomol 116:118–126 Martin JC, Kuvangkadilok C, Peart DH, Lee BTO (1980) Multiple sex determining regions in a group of related
123
Aquat Ecol Chironomus species (Diptera: Chironomidae). Heredity 44:367–382 McKie BG (2004) Disturbance and investment: developmental responses of tropical lotic midges to repeated tube destruction in the juvenile stages. Ecol Entomol 29:457–466 McLachlan AJ (1983) Life history strategies of rainpool dwellers. J Anim Ecol 52:545–562 McLachlan AJ, McLachlan SM (1976) Development of the mud habitat during the filling of two new lakes. Freshwater Biol 6:59–67 McLachlan AJ, Yonow T (1989) Reproductive strategies in rainpool dwellers and the model freshwater insects. Hydrobiologia 171:223–230 Pinder LCV (1995) Biology of eggs and first instar larvae. In: Armitage PD, Cranston PS, Pinder LCV (eds) The Chironomidae: the biology and ecology of non-biting midges. Chapman & Hall, London, UK, pp 87–106 Price PW (1984) Insect ecology. Wiley, New York Rasmussen JB (1985) Effects of density and microdetritus enrichment on the growth of chironomid larvae in a small pond. Can J Fish Aquat Sci 42:1418–1422
123
Renshaw M, Service MW, Birley MH (1993) Density dependent regulation of Aedes cantans (Diptera: Culicidae) in natural and artificial populations. Ecol Entomol 18:223–233 Sankarperumal G, Pandian TJ (1991) Effect of temperature and Chlorella density on growth and metamorphosis of Chironomus circumdatus (Kieffer) (Diptera). Aquat Insects 13:167–177 Schmid PE (1992) Population dynamics and resource utilization by larval Chironomidae (Diptera) in backwater area of the River Danube. Freshwater Biol 28:111–127 Stevens MM (1998) Development and survival of Chironomus tepperi Skuse (Diptera: Chironomidae) at a range of constant temperatures. Aquat Insects 20:181–188 Teng HJ, Apperson CS (2000) Development and survival of immature Aedes albopictus and Aedes triseriatus (Diptera: Culicidae) in the laboratory: effects of density, food and competition on response to temperature. J Med Entomol 37:40–52 Tokeshi M (1995) Random and aggregation analyses of dispersion in an epiphytic chironomid community. Freshwater Biol 34:567–578
