Seven yellow eels (572-643 mm, 318-592 g) Anguilla anguilla (L.) were tagged with surgically implanted radio transmitters (activity circuit, 1.6-1.7 g) and tracked in the Awirs stream, a small (width < 5 m, depth from 0.1 to 1.2 m), densely populated (ca. 250 kg of eel ha −1 ) tributary of the Belgian River Meuse. The eels were positioned daily from late April to mid-August, and their diel activity was studied during twenty four 24-h cycles. During day-time, the eels were resting in rootwads or in crevices inside stone walls or in crevices in between rocks. They became more active in the late afternoon but generally did not leave their residence before sunset, except under overcast weather. Activity peaked during the first part of the night then progressively vanished, and always ended before sunrise. The area exploited during night-time never extended over more than 40 m 2 , except when the eel changed its residence. The intensity and timing of nocturnal activity and the extent of the daily activity area were dependent on water temperature (respectively P < 0.0001, P < 0.05 and P < 0.0005), with eels showing little or no activity when the diurnal temperature did not exceed 13 • C. Eels showed higher agitation under full moon and maintained their activity later in the night (P < 0.05). The eels showed restricted mobility, and occupied small stream areas (from 0.01 to 0.10 ha) in a non sequential mode, except for two fish which were displaced to the River Meuse by a spate in early June and were never recovered. The length and frequency of net daily journeys were higher (P = 0.005) at water temperatures above 16 • C in late May and June, which also corresponded to the period of immigration of eels from the River Meuse. This study thus shows that large yellow eels may adopt a highly sedentary lifestyle in a continental, fast flowing and densely populated environment, even at periods of the year when these stages usually show upstream migrations.

Externally attached telemetry transmitters are unsuitable to tag yellow eels Anguilla anguilla (L.), in streams where they exhibit cryptic life habits and hide in narrow cavities between rocks. We evaluated the adequacy of surgical implantation and closing procedures for tagging eels with biotelemetry transmitters. Epoxy dummy transmitters (18 × 8 mm, 1.6-1.7 g) were implanted in eels anaesthetised with 2-phenoxy-ethanol (0.9 ml l −1 ), through a 20 mm mid ventral incision made in the posterior quarter of their body cavity. The incision was either left open, or closed in different ways: stitches (absorbable or non absorbable suture material) or commercial-grade cyanoacrilate adhesive (Loctite TM ). Fish were stocked in a 4 m 2 flow through tank (15-17 • C), controlled daily for mortality and weekly for evaluating the healing process.No transmitter was expelled over a 12-week period, even in eels with unclosed incisions, of which 50% healed within 28 days (t 50 ). Regardless of the nature of the filament, suturing induced skin and muscle necrosis, caused significantly higher mortality rates (60% after 10 weeks) and paradoxically slowed down the healing rate (40 and 45 d, respectively). Cyanoacrilate suppressed the inflammatory response and granted higher survival rate (90%), but did not permit to speed up the closing process (t 50 = 52 d), as eels actively bit and removed the adhesive within hours. This behaviour was suppressed when we applied a freshly cut fragment of the eel dorsal fin as a biological bandage over the drying cyanoacrilate. The adhesive remained in place for one to two days and permitted to substantially increase the healing rate (t 50 = 15 d). These results substantiate the efficiency of surgery techniques for tagging eels with radio transmitters, at least for units of small weight and bulk.

Cannibalism among sibling dorada Brycon moorei started before the yolk sac was completely absorbed, as soon as oral teeth were completely developed (1·2 mg fish, 21 h after hatching, 39 h after fertilization at 27·0 0·5 C under 12L : 12N). Embryos attacked siblings of equivalent size or slightly larger than themselves, which were incompletely ingested, sucked up to the head and regurgitated (type Ia cannibalism). Two-day-old larvae performed complete ingestion but could not digest the head of their prey, which was regurgitated (type Ib cannibalism). One day later, all cannibals had turned to complete (type II) cannibalism. Type II cannibalism persisted during the larval period (ending c. 144 h after hatching, 26-38 mg fish) and the early juvenile stage (15-30-g fish). The logistics of type II cannibalism (maximum prey to cannibal weight ratio, W p : W C in percent) was modelled as 11·9607 W 0·3429 C (r 2 =0·974, P<0·0001), where W C is the body weight of the cannibal (g), indicating that cannibals had to turn to increasingly smaller prey during their ontogeny. When being offered prey of different sizes, cannibals of all sizes (0·04-27 g) preferred the smallest prey available almost systematically. A shortage of prey of appropriate size caused them to turn to larger prey, and eventually to exert incomplete cannibalism over siblings exceeding the maximum W p : W C ratio. Cannibals could ingest extremely high food rations [R max (% W C )=47·4242 W 0·4002 C ; r 2 =0·906, P=0·0126], and showed extremely fast growth (G [% day 1 ]= 2·5895+0·5194 R-0·0007 R 2 ; r 2 =0·974, P<0·0001). These traits caused cannibalism in dorada to be the earliest and most intense ever reported in fish (95-98% fish cannibalized within the first week, of which c. 40% on the first day). The functional, adaptive and evolutionary implications of early predation and shifts between types of cannibalism in dorada are discussed.

No abstract

No abstract