The two effects did not arrive through the same developmental window, and that is the part with management consequences. The 7% length deficit tracked exposure after hatching [5]. The escape failures were reported for the group that heard motorboats at both stages [9], and the comparison figure pools every fish that heard reef sound during at least one stage [8], so the summary does not separate the two single-exposure groups.
Multiply the two escape numbers to get the share of fish that both reacted and reacted away from the threat. For the reef-sound groups: about 80% responded, about 20% of those went the wrong way, leaving roughly 64% clean escapes. For fish raised under boat noise throughout: 68% responded, 40% of those went the wrong way, leaving about 41% [10]. The gap is around 23 percentage points, a bit more than a third fewer clean escapes [11]. The reported percentages are not two separate problems; they compound on the same animals.
Two design choices do real work. Exposure was assigned by the experimenters across four playback combinations for up to 78 days [3][4], so this is a manipulation rather than a correlation between busy anchorages and small fish. And nothing was played for 24 hours before or during the predator test [6], which rules out the simplest reading: that engine noise masks the cue at the moment of the strike. Whatever changed outlasted the speaker, which is why the paper is framed around carryover [20].
What the write-up does not tell you is how loud and how close. No received sound levels or vessel distances are given [17], and the group percentages arrive without sample sizes [16]. The threat itself was a weight dropped on the water surface, scored against the C-shape-then-burst reaction a spiny chromis normally makes [7]; no survival or mortality outcome is reported [18]. Body length is also a partial measure of size. If shape were unchanged, a fish 7% shorter would carry about 20% less volume [21], but the summary reports length, not mass.
On mechanism, Exeter's Sophie Nedelec offers candidates rather than conclusions. Earlier work found motorboat noise disturbs parental care, including the fanning that oxygenates eggs, which is linked to egg survival [12]. If embryonic exposure sensitises the fish, she suggests, they may react more strongly as juveniles, with stress or disrupted feeding accumulating into poor risk assessment and smaller bodies [13]. Her reading of the result is that early-life noise can disrupt escape behaviour even when the escape happens in quiet water [14], and that limiting motorboat traffic close to reefs during the breeding season may reduce the effect [15].
That recommendation is cheap and, on this evidence, defensible, with one condition attached: the study establishes a hazard from an assigned dose, not the distance at which a real outboard delivers it.
