Alteration of the chemical environment disrupts communication in a freshwater fish

Abstract

It is well established that changes to the chemical environment can impair development, physiology and reproductive biology; by contrast, impacts on communication have not been widely reported. This is surprising given that chemical communication is the most widely used sensory modality in nature, and that variation in the chemical composition of the environment is the rule, not the exception. Here, we show that chemically mediated species recognition in a swordtail fish, Xiphophorus birchmanni, can be hindered by anthropogenic disturbance to the signalling environment. Females have a strong preference for conspecific male chemical cues, yet they hybridize in nature with the congener X. malinche. Wild-caught females showed a strong preference for conspecifics when tested in clean water, but failed to show a preference when tested in stream water subject to sewage effluent and agricultural runoff. We hypothesized that this was due to the interaction between chemical communication systems and humic acid (HA), a ubiquitous, natural product elevated to high levels by anthropogenic processes. When exposed to elevated concentrations of HA, female X. birchmanni again lost their preference for conspecific male chemical cues, while visual mating preferences and motivation to mate were retained. Sub-lethal concentrations of seemingly benign substances can thus have a drastic effect on natural populations through their specific impact on communication systems.

Figure 1

Wild-caught, sexually mature male swordtails. (a) Pure X. birchmanni and X. malinche. (b) X. birchmanni/malinche hybrids collected in the Río Calnali at Chahauco. Hybrids display virtually all possible combinations of male traits and often show traits outside of the range of parental species (Rosenthal et al. 2003).

Figure 2

Association time (s; mean±s.e.) of female X. birchmanni with chemical cues of conspecific (black bars) and heterospecific X. malinche (white bars) males. ‘Río Calnali’ indicates trials conducted in water from the stream; ‘Spring’ represents trials conducted in natural spring water; ‘Tap’ indicated trials conducted in municipal tap water. Asterisks indicate a significant preference for conspecific cues over heterospecific: *p<0.05.

Figure 3

Responses (mean±s.e.) of female X. birchmanni to chemical conspecific (black bars) and heterospecific X. malinche (white bars) male cues. ‘Humic acid’ represents experimental fish that were subjected to a preference trial conducted in 200 mg l⁻¹ HA; hatched-shaded area depicts the one trial conducted in HA. ‘Control’ indicates fish that were consistently tested in water not containing supplemental HA. Bars surrounded by a black box indicate a significant difference (p<0.001) in females' response to conspecific male cues relative to other trials (bars not included in boxed area). Asterisks indicate a significant preference for conspecific cues over heterospecific: *p<0.05; **p<0.01, ***p<0.001. (a) Latency to respond to conspecific and heterospecific odour cues. (b) Females' association time with each cue. (c) The frequency that females were within one body length of conspecific and heterospecific stimuli outflow.

Figure 4

Association time (mean±s.e.) of female X. birchmanni with computer animations of conspecific (black bars) and heterospecific X. malinche (white bars) males. ‘Humic acid’ indicates trials conducted in 200 mg l⁻¹ HA; ‘Control’ represents trials conducted without supplemental humic acid. Asterisks indicate a significant preference for conspecific cues over heterospecific: *p<0.05; **p<0.01.