(a) Experimental animals
Adult D. rerio were from an existing stock at the University of Plymouth. Danio albolineatus were obtained from local aquarists. Single species groups of males and females were kept separate in 10 l glass aquaria (26 ± 1°C; 12 h light : 12 h dark) and fed three times daily ad libitum on flake and Artemia. For embryo collection, males and females of a single species were placed together in a breeding tank and embryos collected immediately post-fertilization. A mixture of individuals from several different breeding groups was used during the course of the experiment to maximize genetic variation among embryos. As the precise time of fertilization was difficult to determine, the two-cell stage of development was used as our initial time point.
(b) Experimental setup
Individual embryos were filmed from the two-cell stage for a period of 24 h using a camera (WATEC 202B/WATEC 221S, WATEC Co. Japan), connected to a VCR (Time-Lapse VCR AG-6730, Panasonic Corp. Japan/Computar 24 h Time-Lapse CTR-3024, CBC Corp., USA). Images were captured at 75× magnification, and were viewed on a monitor (WV-CM1470, Panasonic Corp., Japan) connected to the recorder.
To maintain embryos within the camera frame during filming, they were placed into a well made out of a 5 per cent agarose gel held in a 6.5 cm Ø Petri dish. These wells mimicked a structurally complex substrate where zebrafish would oviposit. The diameter of each well was approximately twice the length of an embryo to avoid growth constraints. The Petri dish was maintained at 28.5 ± 1°C. A continuous flow of water (0.3 ml min−1) to the dish was provided via a peristaltic pump from stock tanks. To minimize disturbance, the water input was >5 cm from the well where the embryo was located.
(c) Experimental protocol
Alarm substance was obtained using the method of Sloman et al. (2008)
. Danio rerio
embryos were exposed to alarm substance obtained from adult D. rerio
and D. albolineatus
embryos to alarm substance from adult D. albolineatus
. The concentrations of alarm substance used here did not negatively affect survival to first feeding.
At the two-cell developmental stage, for embryos exposed to alarm substance (n = 11, D. rerio; n = 10, D. albolineatus), 2 ml of alarm substance (at 28.5°C) was pipetted continuously into a well containing an embryo. The alarm substance was removed from the tank after 30 min by flushing the system using a peristaltic pump (75 ml min−1). In control experiments, 2 ml of system water (at 28.5°C) replaced alarm substance (n = 11, D. rerio; n = 11, D. albolineatus).
The time of appearance of three developmental features were recorded for each embryo: eye primordium (i.e. the first appearance of a clear oval shape of the eye contour in the head region of the developing embryo), initiation of the first muscular contraction and first heartbeat.
(d) Statistical analyses
Data (decimalized timings expressed in minutes) were normally distributed and analysed parametrically using SPSS 15.0 for Windows (SPSS Inc. Chicago, IL, USA). The overall difference in the time it took from the two-cell stage to reach the onset of the eye primordium, the first muscular contraction and the first heartbeat was compared between control embryos and those exposed to alarm substance using a repeated-measures analysis of variance (ANOVA). One-way ANOVAs were then used to test for differences in specific timings between stages (i.e. using each stage as an initial time point and testing for differences in the time taken to reach further stages between control and alarm substance samples).