Song structure, not high-frequency song content, determines high-frequency auditory sensitivity in nine species of New World sparrows (Passeriformes: Emberizidae)

1. The evolution of vocal signals can be constrained by a host of factors including habitat effects on sound propagation, morphology of sound-producing structures, and phylogenetic relationships among species. Here, we asked whether auditory sensitivity over a broad range of frequencies correlates with the spectral content of conspecific vocalizations, or whether it is constrained by the overall structure of vocalizations, habitat effects on sound propagation, or relatedness among species. 2. We studied nine New World sparrows (Passeriformes: Emberizidae) including three open-habitat species, three scrub-like habitat species, and three forest species. For each habitat, one species had pure trilled songs, another had tonal songs, and another had complex songs with tones, trills, and amplitude-modulated buzzes. 3. As predicted by the acoustic adaptation hypothesis, song spectral properties (specifically frequency and entropy) had the highest values in open-habitat species and the lowest values in forest species. 4. Based on our results from song analyses, and the sender-receiver matching hypothesis, we predicted that open-habitat species would be more sensitive to high-frequency sounds compared to forest species. Contrary to this prediction, habitat and high-frequency song content had little effect on audiogram shape. Song type, however, had a strong effect, with species that produce complex songs showing higher sensitivity to high-frequency sounds than all other species. 5. Our results suggest that the use of song frequency by receivers depends on song structure and not necessarily on song spectral content. Therefore, our current understanding of how signal-processing mechanisms should match signal properties appears to be too simple. When thinking about the evolution of signal-processing mechanisms, the multidimensionality of signals, and how the different dimensions can interact, should be considered.

1. 发声信号的演化会受到诸多因素的制约，其中包括生境对声音传播的影响、发声结构的形态特征，以及物种间的系统发育关系。本研究旨在探究两个核心问题：一是宽泛频率范围内的听觉敏感性是否与同种发声信号的频谱特征存在关联；二是听觉敏感性是否受鸣唱整体结构、生境对声音传播的影响，或是物种间的亲缘关系所限制。
2. 我们以9种新大陆雀类（雀形目：鹀科）为研究对象，其中包含3种开阔生境物种、3种灌丛生境物种以及3种森林生境物种。每一生境类群中，各有1个物种拥有纯颤音鸣唱，1个物种拥有调式鸣唱，另有1个物种则具备融合了调式声、颤音与调幅嗡鸣的复杂鸣唱。
3. 正如声学适应假说（acoustic adaptation hypothesis）所预测的那样，鸣唱的频谱特征（具体为频率与熵值）在开阔生境物种中达到最高水平，而在森林生境物种中处于最低水平。
4. 基于鸣唱分析结果与发送者-接收者匹配假说（sender-receiver matching hypothesis），我们曾提出预测：相较于森林生境物种，开阔生境物种对高频声音的敏感性更高。但该预测与实际研究结果相悖：生境类型与鸣唱的高频内容对听图（audiogram）的形状几乎无影响。然而鸣唱类型对听觉敏感性存在显著影响，具备复杂鸣唱的物种对高频声音的敏感性显著高于其余所有物种。
5. 我们的研究结果表明，接收者对鸣唱频率的利用取决于鸣唱结构，而非必然取决于鸣唱的频谱特征。因此，当前学界对于"信号处理机制应如何与信号特征相匹配"的认知似乎过于简化。在探讨信号处理机制的演化过程中，应当考虑信号的多维性，以及不同维度间的相互作用方式。