Breeding ecology of the Little Bittern Ixobrychus minutus in northeast Algeria

Abstract

Capsule This is the first study of the reproductive ecology of the Little Bittern Ixobrychus minutus in North Africa. During the study period, clutch size exhibited a sharp drop concurrent with a steady human encroachment and 30% habitat loss.

Aim To investigate the status, nest-site selection, breeding parameters, chick development and chick diet of Little Bittern in Algeria, under deteriorating habitat conditions

Methods Data were collected during surveys of all major Algerian wetlands (2002–08) and monitoring of breeding at Boussedra, a freshwater marsh in northeast Algeria, during the years 2005 and 2008.

Results Egg-laying spanned the period of 10 May to 30 June with one recorded case of a double brood in early July (n = 30). The mean clutch size in 2005 was 6.2 ± 0.4 sd eggs per nest (n = 6 clutches), which was significantly higher than that of 2008, which amounted to 4.3 ± 0.5 sd eggs per nest (n = 6 clutches), and this decline was associated with increased disturbance and loss of habitat. Overall, clutches suffered from a low rate of predation (17.6%) and displayed a high hatching rate (85.7%).

Conclusion Breeding parameters of the Little Bittern in North Africa are similar to elsewhere in Europe.

The Little Bittern Ixobrychus minutus is a marsh heron that is difficult to survey or observe (Kushlan & Hancock 2005). The status of this species in the western Mediterranean is of serious conservation concern and in need of close monitoring (Marion 1994). Knowledge of the breeding ecology of the species is mainly based on early work (Groebbels 1935, Steinfatt 1935, Wackernagel 1950, Grosskopf & Graszynski 1958, Braschler et al. 1961) but this has recently been updated (Langley 1983, Holmes & Hatchwell 1991, Caballero Soler 1997, Pardo-Cervera et al. 2010, Fazili et al. 2010). There is, however, scant information about the breeding of the Little Bittern in North Africa, increasingly threatened by severe anthropogenic pressure on wetlands (Hollis 1992). The Little Bittern is known to be an area-sensitive species occurring only in marshland fragments larger than 4.0–4.6 ha (Pezzo & Benocci 2001, Benassi et al. 2009), thus the unabated impact of habitat loss and fragmentation is bound to threaten the viability of the North African population (Lande 1987, Venier & Fahrig 1996). The present study aimed to: (1) increase our knowledge of the breeding ecology of this species in the region; and (2) monitor changes in breeding parameters of a population facing a rapid shrinkage and deterioration of its habitat.

METHODS

As part of a study of the ecology of waterbirds of Numidia, northeast Algeria (Fig. 1), a survey of all major wetlands of the region was conducted between 2002 and 2008 (Samraoui & Samraoui 2008, Samraoui et al. 2011). Frequent visits in the region, involving half-monthly transects, were undertaken by between two and four people throughout the breeding period (February–August). The breeding ecology of Little Bittern was investigated in 2005 and 2008 at Boussedra (36°50ʹ45″ N, 7°43ʹ47″ E), an unprotected freshwater marsh of 55 ha that has been used as a landfill since 2003, losing 30% of its former area up to 2011. The wetland is a typical Mediterranean wetland fragment (Battisti et al. 2008, Paracuellos 2008), a remnant of a previously large wetland complex around the town of Annaba, and is now embedded in an urban landscape with surrounding hills gradually giving way to housing estates. The vegetation is dominated by Scirpus maritimus and Lesser Bulrush Typha angustifolia. A dense stand of Tamarisk trees Tamarix gallica is located on the southern edge of the wetland. The site hosts the endangered endemic newt Pleurodeles poireti (Samraoui et al. 2012) and is recognized as an Important Bird Area (Fishpool & Evans 2001), being an important wintering ground and a breeding site for the globally threatened White-headed Duck Oxyura leucocephala and Ferruginous Duck Aythya nyroca (Samraoui & Samraoui 2008). Fieldwork was carried out twice a week with each nest labelled with a distinct code and the height of the nest above water and its diameter measured. Vegetation density (% cover) was visually estimated in a 4 × 4-m square centred on the nest. Eggs were weighed using a Pesola spring balance with an accuracy of 0.1 g and their lengths (L), and breadths (B) were measured to the nearest 0.1 mm using vernier calipers. Egg volume (V, in cm³) was calculated using Hoyt's formula (V = 0.000509 L*B ²) (Hoyt 1979). Chicks were individually marked and their mass, total tarsus, head and bill length (tip of bill to back of skull) were recorded between 1 and 20 days. Because chicks were too mobile after six to eight days, only a few measurements could be made after that age. Hatching dates were used to infer the sequence of egg-laying. Information on diet was gathered by collecting food regurgitated by chicks during handling. A small subset of nests was additionally recorded at distinct wetlands, all located within Numidia, northeast Algeria: Garaet Hadj Tahar (36°51′52.0″N, 7°15′48.0″E), Lac des Oiseaux (36° 46′59.0″N, 8°7′13.0″E), Sidi Achour (36°52′30.00″N, 7°43′23.0″E), Lake Fetzara (36°47′0.0″N, 7°31′0.0″E) and Tonga (36°51′0.0″N, 8°30′0.0″E) (Table 1). With the exception of two nests from Lac des Oiseaux, these nests could not be monitored regularly and thus are not included in analyses.

Figure 1. Location map of Numidia, northeast Algeria.

Table 1. Recorded nests with first-egg date (FED) in Numidian wetlands, Algeria.

Wetland	Year	Nests	Vegetation	FED
Boussedra	2003	2	Typha angustifolia	May
	2005	12	Typha angustifolia	18 May
	2008	18	Typha angustifolia	10 May
Hadj Tahar	2002	2	Phragmites australis	Late May
Lac des Oiseaux	2003	3	Juncus acutus, Rubus ulmifolius	Late May
Sidi Achour	2003	3	Juncus acutus, Tamarix gallica	Late June
Tonga	2003	1	Salix atrocinerea, Typha angustifolia	Late May
	2005	1	Iris pseudocarus, Typha angustifolia	Late July
Fetzara	2005	9	Tamarix gallica, Phragmites australis	June

Data analysis involving one-way anova and regression analyses was carried out using r (R Development Core Team 2009). Mean values are given ± 1 sd.

RESULTS

Status

We recorded a maximum of 18 nests at Boussedra (Table 1). This is a sizable proportion (12–25%) of the total estimated number of breeding pairs ranging between 70 and 150 for the whole Numidian wetland complex and inferred from the area of suitable habitats not covered by the survey (Samraoui & Samraoui 2008). Elsewhere in Algeria, the species may only breed sporadically with very low numbers (Samraoui et al. 2011).

Nest characteristics

At Boussedra, with one exception (a nest placed on a bush made up of Tamarisks and S. maritimus), all nests (n = 23) were located in dense stands of Lesser Bulrush with a mean vegetation height of 244.3 ± 42.1 cm (n = 19 nests measured). Nesting heights differed between years with mean nest height in 2005 (20.1 ± 6.9 cm) being significantly lower than that of 2008 (29.7 ± 10.9 cm) (one-way anova: F _1,17 = 5.1, P < 0.05). Confirmation that Little Bittern nests in Numidia are not exclusively confined to Lesser Bulrush was the finding of nine nests at Lake Fetzara, all located in small bushes of Tamarisks mixed with reedbeds of Phragmites australis bordering canals. Nest characteristics are summarized in Table 2, which also summarizes existing information from other studies in Europe, South Africa and India.

Table 2. Nest characteristics of the Little Bittern in Numidia, Algeria and across its geographical range.

	Present study, Algeria	Langley (1983), South Africa	Holmes & Hatchwell (1991), India	Martínez-Abraín (1994), Spain	Grattini (2003), Italy	Pardo-Cervera et al. (2010), Spain	Kayser (2010), France	Fazili et al. 2010, India
Nest characteristics
External diameter (cm)	20.5 ± 2.8 (n = 20)	28.3 (n = 20)		17.7 ± 2.2 (n = 42)		18.1 ± 2.1 (n = 20)
Nesting height (cm)	25.2 ± 10.2 (n = 19)	69.7 (n = 22)		50.5 ± 22.4 (n = 27)	40 (n = 22)	37.5 ± 14.3 (n = 20)
Vegetation height (cm)	244.3 ± 42.1 (n = 19)
Vegetation density (%)	78.7 ± 21.0 (n = 21)
Water depth (cm)	58.6 ± 15.1 (n = 21)	Range: 40–120		31.7 ± 11.9 (n = 43)		35.9 ± 14.9 (20)
Breeding parameters
Clutch size	5.2 ± 1.0	3.5	4.0 ± 0.4	5.5 ± 0.9	4.7	4.0 ± 0.9		5.6 ± 0.7
Range	4–7	3–4	3–5	3–7	NA	1–6		5–7
n	14	14	16	43	16	43		17
Hatching rate (%)	85.7	62.3	70–71	92.8	84	46		68.8
Diet
Fish	66.7%	71.9%	77.8%	78.8%		38.7%	29.4%
Amphibians	20.0%	12.3%	7.4%	12.1%		7.1%	43.9%
Insects	13.4%	Present	13.6%	9.1%		27.7%	22.5%
Crustacea						23.9%
Leeches							1.4%
Spiders							1.4%
Lizards		1.8%					0.5%
Birds							0.9%
Other		14.0%	1.0%			2.6%
No. prey	15	57	81				214

Phenology

Egg-laying was carried out from 10 May to 30 June with, in 2005, a case of a second brood recorded at Boussedra in one nest where the onset of a second clutch (11 July) occurred one week after the last chick of the first brood had fledged.

Egg characteristics

At Boussedra, the mean egg length was 35.1 ± 1.3 mm (n = 24 clutches), the mean egg breadth was 25.7 ± 0.7 mm (n = 24 clutches), and the mean egg volume was 11.8 ± 0.9 cm³ (n = 24 clutches). The mean egg weight for 2005 was 12.7 ± 0.6 g (n = 11 clutches). Egg volume was similar between years.

Clutch size and hatching success

The mean clutch size at Boussedra in 2005 was 6.2 ± 0.4 (n = 6 clutches) whereas in 2008, it decreased to 4.3 ± 0.5 (n = 6 clutches), exhibiting a highly significant drop between the two years (one-way anova: F _1,10 = 46.5, P < 0.0001). Taking into account two complete clutches of 5 eggs recorded at Lac des Oiseaux in 2003, the overall mean clutch size recorded in Numidia amounted to 5.2 ± 1.0 (n = 14 clutches) (Table 2).

Over the two-year study period, predation was responsible for 17.6% of failed clutches (n = 17) and the hatching rates for successful clutches (with at least one recorded hatchling) were 81.1% (n = 6) and 92.3% (n = 6) for 2005 and 2008, respectively. Chick productivity was difficult to assess because chicks were prone to leave the nest a few days after hatching. The survival rate of nestlings was assumed to be quite high (>90%) because no dead chick was ever recorded in or around the nest.

Nestling growth

Data of chick growth were obtained from 13 chicks, from 5 nests monitored in 2005. Growth rates were best modelled by a linear regression for chicks between 5 and 12 days of age with mass gain (g d⁻¹), tarsus (mm d⁻¹) and head and bill (mm d⁻¹) growth averaging 10.2 ± 0.5, 2.9 ± 0.2 and 4.1 ± 0.2 (r ² ≥ 0.95, P < 0.05, n = 13 chicks from 5 nests, for all 3 growth rates), respectively. The unequal contribution of each nest (pseudoreplication) was tolerated here because of the limited sample of nests.

Diet

A total of 15 prey were identified from the 5 pellets regurgitated by nestlings from 5 distinct nests collected between June and August 2005. Prey consisted of aquatic insects and small vertebrates (Table 2): Mosquitofish Gambusia holbrooki (66.7%), Sahara frog Pelophylax saharicus (20.0%), notonectid Anisops sardea (6.7%) and adult damselfly (Zygoptera) (6.7%).

Brood parasitism

One nest was discovered at Lake Tonga on 6 June 2003 containing three eggs of Little Bittern and one egg of Moorhen Gallinula chloropus. The nest was not subsequently monitored and the fate of the eggs is thus unknown. Conspecific brood parasitism may be involved in the case of a clutch of seven eggs found at Boussedra in 2008. Six chicks successfully fledged out of this nest.

DISCUSSION

Prior to the present study, little was known of the reproductive ecology of the Little Bittern in North Africa. In the study area, the species nested in marshes or freshwater ponds with dense stands of emergent vegetation. Thus, the pattern of nest-site selection recorded was similar to studies carried out elsewhere: Little Bittern nests are predominantly associated with reedbeds but the species is also known to select low trees and bushes in marshes (Boileau & Barbier 1997, Delelis & Boin 2006).

The onset of egg-laying in Numidia was found to occur relatively earlier than was recorded in the Ebro delta, Spain (Pardo-Cervera et al. 2010) but the peak of breeding (June) is similar for both regions. Egg-laying was spread from 10 May to 30 June with mean clutches ranging from 4.3 to 6.2. Previous data on Little Bittern from Algeria gave clutches with a range of three to four eggs laid in May (Heim de Balsac & Mayaud 1962). The Little Bittern was also found in the present study to be occasionally double-brooded in Numidia as was found in Kashmir (Holmes & Hatchwell 1991). In the Least Bittern Ixobrychus exilis in North America, there is a seasonal increase in clutch size (late clutches undoubtedly included some second clutches), suggesting a similar pattern in trophic resources (Weller 1961). Although rare, clutches of seven eggs were noted in the present study but larger clutches (up to eight) were previously recorded by Witherby (1943), who suggested that conspecific brood parasitism might be involved. The mean clutch size of 3.5 eggs for I. m. payesi in South Africa (Langley 1983) is smaller than that recorded for the nominal species (Table 2), a probable consequence of a range closer to the Equator (Voisin 1991). No other apparent latitudinal variation of clutch size seems to be present in the Little Bittern as has been shown in the Least Bittern (Post 1998).

While it is not clear what may be the real cause of the marked decline in the mean clutch size between 2005 and 2008, and bearing in mind that we only have two years' data, habitat deterioration, shrinking by 30%, is a plausible explanation. In the study area, extensive studies of colonial herons (Ardea ibis, Egretta garzetta, Ardeola ralloides and Nycticorax nycticorax) have shown that larger clutch sizes are consistently associated with larger wetlands (Samraoui Chenafi 2009). Disturbance by trucks and decreasing foraging areas were also likely to have affected breeding parameters such as clutch size and productivity.

With a mean mass gain per day of 10.2 g, the development of chicks in Numidia was similar to that reported previously (Holmes & Hatchwell 1991), with chicks being 10–11 g at hatching and attaining 100–125 g before fledging. For larger herons, the mean mass gain per day in Numidia ranged from 14.5 g for the Cattle Egret (Samraoui et al. 2007) to 49.6 g for the Grey Heron (R. Nedjah et al., unpubl. data).

Predation is an important evolutionary force that strongly influences the breeding success of birds (Post 1990, Martin 1993). Compared with other studies (Pardo-Cervera et al. 2010) carried out in rice fields and in the presence of an abundant population of rats Rattus rattus and R. norvegicus, predation on Little Bittern nests in Numidia appears to be relatively low. Studies of the Least Bittern have indicated that it may benefit from close association with a larger and much more aggressive species, the Boat-tailed Grackle Quiscalus major (Post & Seals 1993, Post 1998). In Numidia, nests of Little Bittern are often found in close proximity to Great Reed Warbler Acrocephalus arundinaceus nests but further investigations are needed to clarify the true nature of this association.

Preliminary results of diet of the Little Bittern in Algeria indicate that diet composition is similar to previous studies (Langley 1983, Holmes & Hatchwell 1991, Martínez-Abraín 1994). While there is considerable variation in the relative importance of fish across different regions (Kayser 2010, Pardo-Cervera et al. 2010), there is remarkable convergence on the importance of three major taxa (fish, anurans and aquatic insects) making up the staple diet of the Little Bittern. Reliance on such a small set of prey implies that the conservation of the Little Bittern in Algeria is tightly linked to the fate of its aquatic habitats, which are in urgent need of effective protection and management (Riservato et al. 2009, Samraoui et al. 2012).

In the third case ever recorded of interspecific brood parasitism involving the Little Bittern, the Common Moorhen was once more recorded as the parasitic species (David et al. 2005, Pardo-Cervera et al. 2010). The fate of the brood in the present study unfortunately remained unknown but cases of successful hatching of parasitic Moorhen eggs are known (Post & Seals 1989, Ueda 1993). The Common Moorhen is known to exhibit a wide range of breeding strategies involving both conspecific (Gibbons 1986, McRae 1998) and interspecific (Forman 2003) brood parasitism. Although records of interspecific brood parasitism of the genus Ixobrychus by the Moorhen are relatively rare, the geographic range covered spans the Old and the New World suggesting that the phenomenon may be under-recorded. Egg recognition and ejection of foreign eggs do not seem to have evolved in Least or Little Bitterns (David et al. 2005) because experimental manipulations indicated that attachment to the nest was stronger than that expressed for the eggs (Weller 1961).

It is well established that habitat loss and fragmentation are threatening many species, causing rapid declines and extinctions (Saunders et al. 1991). Focus of research on the negative effects of habitat fragmentation has been greater on terrestrial ecosystems (Van Dorp & Opdam 1987, Villard et al. 1999), with much less attention devoted to wetlands (Naugle et al. 2001), in particular Mediterranean wetlands (Battisti et al. 2008, Paracuellos 2006). Nesting birds have been shown to be particularly sensitive to progressive fragmentation because they may exploit differentially distinct patches (for breeding, foraging, and dispersal) within a heterogeneous landscape (Paracuello 2008). Locally, despite conservation gains represented by an increasing number of protected wetlands, dam construction and human encroachment are fast destroying reed marshes and other important wetlands (Samraoui et al. 1992), thus threatening habitat specialists and area-sensitive species like the Little Bittern (Benassi et al. 2009). Between 2003 and 2008, Boussedra lost 30% of its area and, if this rate continues unchecked, it will probably face the same fate as other Numidian wetlands which have disappeared over the last two decades (de Bélair & Samraoui 1994). The rate of habitat erosion is alarming because many species rely on the close connectivity of local and regional wetlands (Boucheker et al. 2011, Baaloudj et al. 2012) and we expect that many species cannot indefinitely sustain the loss of key habitats (Semlitsch & Bodie 1998) that may prove critical to ensure regional persistence through local extinction–colonization dynamics (Taylor et al. 1993, Hanski 1998). Monitoring reed bed species which have been shown to be sensitive to landscape features like area, edge effects and fragmentation (Báldi 1999, Paracuellos 2006), and information on demographic parameters such as breeding success in rapidly deteriorating habitats may inform how waterbirds respond to such environmental changes. Such critical knowledge may then guide ongoing conservation efforts to mitigate the progressive erosion of wetland biodiversity (Tscharntke 1992, Gibbs 2000).

ACKNOWLEDGEMENTS

We are most grateful to Will Cresswell and two anonymous referees for their helpful comments. The work was supported by the Algerian Ministère de l'Enseignement Supérieur et de la Recherche Scientifique (DGRSDT/MESRS) and DSFP, King Saud University, Saudi Arabia.