, Volume 137, Issue 4, pp 617–620

Founder population size and number of source populations enhance colonization success in waterstriders


  • Petri Ahlroth
    • Department of Biological and Environmental Science University of Jyväskylä
  • Rauno V. Alatalo
    • Department of Biological and Environmental Science University of Jyväskylä
  • Anne Holopainen
    • Department of Biological and Environmental Science University of Jyväskylä
  • Tomi Kumpulainen
    • Department of Biological and Environmental Science University of Jyväskylä
    • Department of Biological and Environmental Science University of Jyväskylä
Conservation Ecology

DOI: 10.1007/s00442-003-1344-y

Cite this article as:
Ahlroth, P., Alatalo, R.V., Holopainen, A. et al. Oecologia (2003) 137: 617. doi:10.1007/s00442-003-1344-y


Understanding the factors that underlie colonization success is crucial both for ecological theory and conservation practices. The most effective way to assess colonization ability is to introduce experimentally different sets of individuals in empty patches of suitable habitat and to monitor the outcome. We translocated mated female waterstriders, Aquarius najas, into 90 streams that were not currently inhabited by the species. We manipulated sizes of propagules (from 2 to 16 mated females) and numbers of origin populations (one or two). Three origin populations were genetically different from each other, but they were less than 150 km from the streams of translocation. The results demonstrate clearly that both the larger propagule size and the high number of source populations have positive effects on the probability of colonizing a new stream. Thus, in addition to the stochastic factors related to the propagule size it may be essential to consider also the diversity of genetic origin for colonization success.


Aquarius najasFinlandGerridaePropagule sizeStreams


Introduction or reintroduction of a species is an important practice in animal conservation. Exploring the causes that determine colonization success and the risk of extinction have become increasingly important in population ecology and conservation biology (Lande 1988; Griffith et al. 1989). One of the most important factors that determines whether the colonization will be successful is the number of individuals that constitute the propagule (the number of colonizing individuals), as verified by studies from many different taxa (Ebenhard 1989; Griffith et al. 1989; Schoener and Spiller 1995; Veltman et al. 1996; Berggren 2001; but for an exception see Schoener and Schoener 1983). Increased founder population size may increase the probability of population persistence or colonization success by decreasing the effects of demographic stochasticity, decreasing the effects of loss of genetic diversity by drift, and decreasing the amount of inbreeding depression. Other factors affecting colonization success include the size and the quality of the patch (Schoener and Schoener 1983) and in some cases the absence of predators (Schoener and Spiller 1995).

The relationship between colonization and population genetic structure is complex. Mathematical models suggest that colonization success depends on the number and origin of individuals founding the populations (McCauley 1991). The effects of inbreeding depression and genetic drift, which did not directly affect colonization success, have been investigated previously (Newman and Pilson 1997; Saccheri et al.1998). However, the empirical studies focusing on the effects of genetic diversity on colonization success and population persistence are missing. Mathematical models suggest that increased genetic diversity may increase the probability of population persistence by increasing the probability that at least some genotypes are adapted to the new habitat (McCauley 1991). To our knowledge, no experiments have been carried out to study both the effects of propagule size and number of source populations on colonization success at the same time. To disentangle these effects it is necessary to control both population size and number of origin (number of genetically distinct populations). In this study, the stream-inhabiting waterstrider, Aquarius najas, was experimentally introduced in streams that were previously devoid of the species.

Materials and methods

The species

The waterstrider is ideally suited to experimental colonization studies. The species is widely distributed in Europe, from southern Fennoscandia to northern Africa (Vepsäläinen 1973). In Finland, it occurs in the south-eastern parts of the country (Vepsäläinen 1973; Lammes and Rinne 1990). The species is wingless in Finland, which decreases the possibility that the released individuals would escape from the study streams (Huldén 1979; Ahlroth et al. 1999). Thus, it occurs as highly isolated populations only in stream habitats, isolated by large lakes or terrestrial habitats. Today, the species is considered to be near threatened in Finland (Rassi et al. 2001).

A. najas, which has only one generation per year, is a large waterstrider species, 11–17 mm in length (Linnavuori 1966; Huldén 1979) and most reproductive females are territorial (Vepsäläinen and Nummelin 1985). One week after copulation females go deep underwater (up to 10 cm) and lay several egg-clusters on stones and on the leaves of aquatic plants (Huldén 1979). Under laboratory conditions, the species is capable of producing up to 400 eggs (average 120 eggs), but the variation between females is large (Ahroth 1999). The eggs hatch during May–June (Huldén 1979). The nymphs go through at least five nymphal stages before they grow to adulthood (Vepsäläinen and Krajewski 1986). Reproducing individuals die after the reproduction season and the surviving offspring reproduce during the next spring. The matured waterstrider offspring overwinter in terrestrial sites such as under rocks or moss in the ground layer of vegetation (Linnavuori 1966; Huldén 1979). Overwintering mortality of A. najas is high, varying from 58% to 96% (Huldén 1979; Ahlroth 1999).

Study areas

At the end of May 1998, waterstriders were caught from three large populations: Jäppilä (J, 62°16′N, 27°33′E), Hankasalmi (H, 62°22′N, 26°32′E) and Leivonmäki (L, 61°59′N, 25°59′E). Each population was estimated to include at least 10,000 mature individuals. They are located in central Finland and separated by 50–80 km from each other. Thus climatically they have experienced similar conditions, but they belong to different watersheds that are not closely connected. We selected the three central Finland populations out of 17 Finnish populations whose genetic differences had been previously studied (Ahlroth 1999). Genetic differences were studied by random amplified polymorphic DNA (RAPD) technique and no clear dependence between genetic identity and distance between populations emerged. The three selected central Finland populations differed significantly from each other, since in a genetic tree they were placed in distant branches all of them having closer neighbors from geographically more distant populations (Ahlroth 1999).

From each stream we collected 216 females and 216 males and kept the sexes in separate buckets. Each bucket was filled with moist Sphagnum moss. The waterstriders were transferred to the laboratory, where each female was mated with only one male from the same population. Presumably some females had already been mated before collection, but they would have been mated with a male with the same origin. The interval between collection and mating in the laboratory was only 1 day. After mating, females were kept 1–2 days in their own buckets containing only individuals from the same origin before introducing them into the vacant streams.

In the experiment, founder propagule size and number of source populations were manipulated by introducing 2, 4, 6, 8 or 16 mated females from either one or two original populations. Firstly, we inventoried 135 possible streams, of which we selected 90 most similar to the species’ natural habitat according to our experience. At the least, small perch (Perca fluviatilis), roach (Rutulus rutilus), bleak (Albornus albornus) or minnow (Phoxinus phoximus) occurred in most of the streams. In early June, a total of 648 females was released into 90 different streams devoid of the species. The native population, the location of the new patch and the day of release were recorded for each female. Since males were not introduced, and the species is univoltine, outbreeding between different populations was prevented during the study period of only one generation. Each treatment propagule was introduced in nine streams that were randomly chosen among the 90 chosen for the colonization experiment. Thus, there were three replicates from J, H, and L for each propagule size in the single origin situations, and three replicates of each combination of populations JH, JL and HL (half of the founder females from both populations) in the situations where the founder propagule of 2–16 mated females included females from two populations. The streams for colonization were located in a relatively small area (60 ×70 km) about 40 km to the north of the northern range of the current natural distribution area in Central Finland and less than 150 km from the sites of origin.

For every stream we measured the mean width (range 0.70–7.72 m), maximum depth (range 15–110 cm) and mean current speed (at three sites per stream) (range 0.05–0.85 m/s) a week after the release of females. The original streams mean width, maximum depth and mean current speed was within the ranges of the introduced streams. These data were collected to control for the quality of the habitats, and all three measures of each variables at each stream were averaged before logistic regression analysis.

Monitoring colonization success

All 90 streams were systematically searched twice for mature offspring of the released female waterstriders. The searching was carried out in the middle of August and in early September. All nymphs would have matured before the first census. In each stream, mature offspring were searched by walking 300 m upstream, and 300 m downstream from the site of release. Seasonal movements of A. najas are typically of only a few meters and they rarely move more than 150 m during the summer (Ahlroth 1999). We caught all the observed A. najas individuals from the stream using insect nets. The initial colonization was considered to be successful if we found mature offspring during either or both of the two visits.


Among the 90 streams, mature offspring were found in 20, while the number of observed individuals varied from 1 to 132 with a median of 9 (in 75% of cases >5 individuals). Both the size of the founder population and the number of source populations significantly affected colonization success according to a logistic regression model (Fig. 1). Thus, the likelihood of successful colonization increased with the founder population size, but for a given propagule size the colonization success was higher for founders from two different origins in comparison to only one source population.
Fig. 1.

Colonization success of introduced waterstriders, Aquarius najas, was highest among those populations that originated from two different source populations (χ2=4.15, df=1, P=0.042) and that were established by large propagule sizes (χ2=5.22, df=1, P=0.022). The lines indicate the predictions from the logistic regression model (χ2=9.5, P=0.009); for each treatment the proportion of streams with successful maturation by the offspring is based on nine replicates. The symbols indicate number of source populations: individuals originated from one source population (square) and from two source populations (dot)

The identity of the source population had no effect on colonization success. Among the translocations with only one source population, each of the three source populations gave rise to two successful introductions. Among the characteristics of the river of release, stream depth had some negative effects on colonization success. In a logistic regression model with treatment type, identity of population origin and the stream characteristics as independent variables, significant positive effects emerged for number of origins (χ2=4.95, P=0.026), for number of female founders (χ2=5.35, P=0.021) and for stream depth (χ2=4.52, P=0.031; for this model χ2=14.28, df=3, P=0.025). Thus, even if increasing stream depth increased the likelihood of offspring survival, the effects of both treatments remained significant in the model.


Propagule size

A major impediment to successful colonization is random extinction before the viable population size has been established. Environmental, ecological and genetic stochasticity can expose small founder populations to extinction (Shaffer and Samson 1985; Schoener and Spiller 1987; Harrison et al. 1988; McCauley 1991; Hanski et al. 1995a, 1995b; Schoener and Spiller 1995; Newman and Pilson 1997; Saccheri et al.1998). Populations are usually founded by very small propagules, and in the initial stages of colonization these populations therefore run a high risk of extinction due to stochasticity. Our results suggest that both the increasing size of propagule and the number of source populations improve colonization success. The result with regard to the founder population size is in accordance with some earlier studies on birds and mammals (Griffith et al. 1989; Veltman et al.1996) and insects (Berggren 2001). These studies indicate that increasing founder population size enhances population persistence, whereas in a study on lizards, the founder population size turned out not to be a significant predictor of population persistence time (Schoener and Schoener 1983). In the case of lizards, longevity and other features in lifespan may explain the contradictory observation. Among insects the single chance to breed may even strengthen the significance of environmental stochasticity.

Quality of environment

Other factors of importance for successful colonization include the suitability of the habitat in the target patch for the founders. Colonization success is always dependent on the quality of the environment, as exemplified by the depth of stream in our case. In waterstriders, the hatching success of eggs may increase with increasing stream depth and in that way the colonization success might be higher. In previous studies with lizards (Schoener and Schoener 1983) and with butterflies (Kuussaari et al. 1996) habitat features have been important for facilitating colonization. For example, Schoener and Schoener (1983) found that time to extinction of lizards increased with increasing island area indicating habitat quality effects on colonization success. High risk of predation has also been found to decrease colonization success. For example, colonization success of spiders was greater on islands without than with lizards (Schoener and Spiller 1995).

Number of source populations

Our results suggest that the number of source populations among colonizing individuals (and thus probably also the diversity of different genotypes) contributes significantly to the colonization success of founder populations. In our study, we examined the initial colonization success before any interbreeding could occur between the source populations. This allowed us to evaluate the importance of genetic diversity on the probability that new colonists would have suitable genotypes allowing them to survive in the new habitat. We suggest that the initial probability of a population to persist in a new environment is dependent upon the amount of genetic variation available, since the greater diversity among colonizing individuals (with two origins) makes it more probable that at least some of the individuals adapt to the new environment.

The amount of genetic variation may also be essential for the long-term survival of the population, although negative effects due to outbreeding depression may arise in future generations if populations of origin are too distant (e.g., Leberg 1993). Outbreeding depression may take place due to a break up of co-adapted gene complexes. On the other hand, if individuals in their original population are inbred, mating between individuals from different origins could result in offspring that enjoy a fitness advantage from hybrid vigor as was demonstrated experimentally with Daphnia magna in rock pools (Ebert et al. 2002). Given the large sizes of the original populations, our founders are not likely to suffer inbreeding problems. In general, inbreeding is another common factor causing problems for small populations (Saccheri et al. 1998; Madsen et al. 1999).

Conservation implications

In Finland, A. najas is considered a near threatened species (Rassi et al. 2001), the main reason being that the quality of aquatic habitats has changed during recent decades due to pollution, digging, dredging and regulation of water levels. Habitat destruction is also causing endangerment and extinctions among other aquatic species (e.g., Rassi et al. 2001; Korkeamäki and Suhonen 2002). In such situations, reliable estimates of minimum propagule sizes to establish viable populations are also needed. Experimental introductions can be invaluable for the management of vulnerable species, which usually have a low colonization success rate (Griffith et al. 1989). However, the critical propagule size that allows successful colonization may vary markedly from species to species. For example, species with higher dispersal tendencies or better survival propensities may provide opposing results with higher or smaller critical propagule sizes, respectively. However, there is an urgent need to accomplish more studies of this kind for applications in nature conservation. Understanding the connections between lifespan, dispersal and extinction risks could allow general theories that provide shortcuts for management of any new, previously unstudied, species.


We thank trainees (Raimo, Mirka) and students (Irma, Susanna, Tero, Jarno, Juha, Marjo) who have helped us during the field season. Katri Kärkkäinen and Mervi Ahlroth provided valuable comments on the manuscript. This work was financed by the Academy of Finland (in two projects: “Local adaptations in small populations” and FIBRE-project: “Viability of populations, assessment of biodiversity and conservation value”) grants to Rauno V. Alatalo. Anne Holopainen was funded by The Entomological Society of Finland and Kuopion Luonnon Ystäväin Yhdistys ry, and Betty Väänäsen rahasto.

Copyright information

© Springer-Verlag 2003