We used adult and young individuals of T. saltator. Adults were collected on a sandy beach in southern Tuscany near the Albegna River mouth (Grosseto, Italy) during the last quarter or new moon. The sea–land direction of the Y-axis (perpendicular to the shoreline) is 268–88°.
The sandhoppers were transported to the laboratory and transferred to Plexiglas containers with wet sand. They were fed weekly with dry fish food on blotting paper. The animals were subjected to an artificial light cycle corresponding in phase and duration to the natural one. There was no artificial dusk or dawn, nor an artificial moon. Daytime illumination was provided by neon tubes (light intensity 65
The experiments were conducted in a dark room in the laboratory from late spring to early autumn 1996–1998 and 2000, in coincidence with a natural moon phase ranging from 94 to 99% (i.e. nearly full moon
). The moon's azimuthal variation during the various experiments is represented outside the circular distributions. Releases were carried out using an experimental apparatus that previously allowed us to reproduce sun compass orientation under artificial light conditions that was comparable to the orientation under the natural sun and sky (; Ugolini et al. 1998
; Ugolini & Castellini 2004
). In short, the apparatus consisted of a circular plate of white Plexiglas (diameter 80
cm) with a central part of transparent Plexiglas (29
cm diameter). As there is no inter-individual influence on the directional choice in sandhoppers (Scapini et al. 1981
), 8–12 individuals at a time were released in a transparent Plexiglas bowl (18
cm diameter) located on the transparent plate, which allowed vision of the animals and the recording of their directions from below. A single direction per individual was recorded 1–2
min after release by means of a video camera underneath the bowl or by directly reading the direction on a goniometer beneath the bowl. A hemispherical dome of white glazed Plexiglas (diameter 80
cm) covered the circular plate. An optical fibre bundle (8
mm diameter) with a negative lens (focal length 11.7
mm diameter) was placed through the central transparent portion of the dish so that it stuck into the bowl to 28
mm. Its function was to illuminate the inner part of the dome (=artificial sky). A second optical fibre bundle (=artificial moon, 4.5
mm diameter) was inserted into the wall of the dome from the outside, at a height corresponding to around 45° from the plane containing the bowl and the animals. The fibre bundles were lit with two variable intensity illuminators (Schott KL 1500). In some releases a black glazed Plexiglas circular plate was added to the floor of the apparatus to reduce the artificial sky intensity. Some tests were carried out using a black Plexiglas dome. The azimuths of the artificial ‘moon’ or ‘sun’ corresponded to those of the natural ones at the moment of the releases (±5°).
Figure 1 Experimental apparatus used to reproduce astronomical orientation under artificial light conditions. (a) external view: a, white Plexiglas dome (80cm in diameter); b, Plexiglas plate; c, optical fibre bundle used to simulate astronomical orientating (more ...)
The different light intensities used in the experiments were measured with a radiometer (Graseby S370, with a silicon photometer and a flat response between 350 and 1100
nm). The light distribution inside the dome was uniform, without asymmetries on the horizontal plane that could have influenced the directional choices of the animals (for a description of the distribution of light intensity inside the dome see Ugolini et al. 1998
). Because of the difference in the emitting surfaces, the artificial moon (or sun) was far more brilliant than the artificial sky.
In the previous study (Ugolini et al. 1998
), a set of filters was placed in front of the radiometer sensor to better compare the sandhoppers' behaviour with their spectral sensitivity curve (see Ugolini et al. 1996
for a preliminary investigation). For the present experiment, we decided not to use any filters in order to render the experimental conditions easier to reproduce. Moreover, we carried out a new set of measures with a different geometry: the sky intensity was measured while the artificial ‘moon’ was screened, and the intensity of the artificial moon was measured while the sensor head was pointed directly at it.
About one month after being collected, adult females were allowed to mate with males of the same population. The young sandhoppers were kept in the laboratory in the same conditions as the adults, i.e. without any possibility of seeing the natural sky and sun. We only tested inexpert young sandhoppers (born in the laboratory and thus lacking experience in the wild), exposed for the first time to the artificial sky and sun at the time of release, 15–30 days after hatching. They had never been exposed to the natural sky, sun or moon.
The statistical analysis was carried out with the methods described by Batschelet (1981)
. For each set of directions recorded in each release, the length and mean angle of the mean vector were calculated. The uniformity of the distributions was evaluated by Rao's test (p
>0.05). The bimodality was assessed by the method of doubling the angles (Batschelet 1981
), even though it does not provide the best description of the distribution of the directions.
For each release, we also calculated the seaward Y
axis tendency (YSc), equal to the homeward component: YSc=r
), where r
is the mean vector length, YS is the seaward direction of the home beach (theoretically expected for perfect orientation), α
is the mean angle of the resulting vector (see Batschelet 1981
, p. 42). In the case of bimodality, the YSc was calculated on the basis of the nearest α
to the expected seaward direction.
The influence of the different light intensities on orientation was evaluated by comparing the various sets of YSc: sets with the number of releases n
≥3 were compared by the Kruskal–Wallis one-way non-parametric analysis of variance, those with the number of releases n
=2 by the Wilcoxon Mann–Whitney test for unpaired data (see Siegel & Castellan 1988