Agalychnis callidryas and A. spurrelli egg masses

Syntopic populations of species provide a unique opportunity to examine divergence of adaptive traits that suggest character displacement when compared with allopatric populations of those species. Past work on syntopic populations of tree frogs, Agalychnis callidryas and A spurrelli documented differences in embryonic risk responses. Due to a serendipitous opportunity, we sampled egg masses from these species oviposited during a breeding aggregation at a forested wetland during July 2017 in the Puntarenas Province of Costa Rica. We collected data on egg masses of the two species including oviposition site and evidence of predation. Both species oviposited on vegetation around the perimeter of the wetland. Additionally, we found that A. callidryas, the species that was previously documented as hatching quickly in response to snake predation took advantage of canopy vegetation overhanging the wetland for oviposition, sites more available to arboreal snakes. This canopy vegetation was not used by A spurrelli. Conversely, A spurrelli, the species shown to respond more strongly to submergence by hatching early, oviposited on vegetation directly emerging from the wetland, sites not used by A callidryas. This divergence of reproductive and ecological traits is not seen in allopatric populations of the same species. These findings suggest that character displacement in oviposition site may have caused differences in selection pressure on embryonic behavior.

Methods
We collected data from Agalychnis callidryas (red-eyed tree frog) and A. spurrelli (gliding tree frog) egg masses in a forested wetland on the Osa Peninsula, Corcovado National Park, Puntarenas Province, Costa Rica during July 2017. The main body of the wetland (08° 28.530’ N, 083° 35.566’ W) was approximately 45 x 20 m with a narrow and shallow area of wetland extending an additional 20 m to the west. The two species of Agalychnis had initiated reproduction in response to intense rainfall. Sampling lasted 6 days: hatching in A. callidryas can occur in as early as 4 days and generally lasts 5-8 days in both species.

 Information on timing, oviposition sites, and predation (egg mass general development, size, oviposition site including substrate, orientation, location in respect to canopy and water surface, and evidence of predation) was collected in the field. We searched the wetland for egg masses on plants in and directly around the wetland, from the water surface up to approximately 3 m above. We would not have been able to reliably see egg masses higher in the canopy and literature suggests that most eggs of A. spurrelli occur lower than 3 m in height (Scott and Starrett 1974). All egg masses we found were within 1.6 m of the water’s surface. We temporarily marked each mass location and visited daily to count embryos and document evidence of predation, disease, or developmental abnormalities. We documented predation when there was direct evidence (sections of embryos and envelopes missing, popped eggs) or > 10 embryo change in count from one day to the next if this occurred early in development (< 4 days based on Cohen et al. 2016). We categorized type of predation as none, missing volume (likely snake predation), envelope remaining but embryo missing (if early in development) or popped (likely invertebrate predation). Masses that were missing embryos but had the envelope after day four of development were assumed to have started hatching.

 We calculated additional variables including the volume, embryo density, and oviposition date for each egg mass based on development at time of detection. We calculated the volume of egg masses based on a spheroid ((4/3*pi*a*b*c)/2 = volume where a, b, and c are the radii for length, width and depth): half the spheroid on leaves or relatively flat surfaces, and the full spheroid for masses surrounding twigs or vines. We calculated density of embryos in an egg mass (maximum number of embryos counted/mass volume) only for egg masses that showed no sign of predation via missing sections of jelly when we initially started counting embryos. An estimated day of oviposition was calculated for those masses we did not find immediately based on subtracting 7, the median time to hatching in both species (Savage 2002), from the hatching date. There was a large rain event on the evening before the last day of sampling and A. callidryas responded with a new flush of laying. To not overinflate confidence in our statistical tests, we did not include egg masses oviposited on the last day in the analyses of oviposition date.

This work was approved by the Texas Tech University Animal Care and Use Committee (17063-07) and covered under research permit INV-ACOSA-053-17 to A. Vega and was conducted in accordance with relevant institutional and national guidelines.

物种的同域种群（syntopic populations）为研究适应性性状的分化提供了独特契机——相较于该物种的异域种群（allopatric populations），这种分化可反映性状替换现象。既往针对红眼树蛙（Agalychnis callidryas）与滑翔树蛙（A. spurrelli）的同域种群研究，已证实二者胚胎风险响应存在差异。得益于一次偶然契机，我们于2017年7月在哥斯达黎加蓬塔雷纳斯省的一处森林湿地，采集了这两个物种在繁殖聚集期所产的卵团。我们收集了两个物种卵团的相关数据，包括产卵位点与捕食证据。两个物种均将卵产于湿地周边的植被上。此外我们发现，此前被证实会针对蛇类捕食快速孵化的红眼树蛙，会选择湿地上方的冠层植被作为产卵位点——这类位点更易被树栖蛇类接触。而滑翔树蛙并未利用这类冠层植被。与之相对，被证实会因遭遇淹水而更早孵化的滑翔树蛙，则将卵产于直接从湿地中长出的植被上，这类位点并非红眼树蛙的产卵选择。这种繁殖与生态性状的分化，在上述物种的异域种群中并未出现。本研究结果表明，产卵位点的性状替换可能导致了胚胎行为所受选择压力的差异。

研究方法
2017年7月，我们于哥斯达黎加蓬塔雷纳斯省科科瓦多国家公园奥萨半岛的一处森林湿地中，采集了红眼树蛙（Agalychnis callidryas，又名红眼树蛙）与滑翔树蛙（A. spurrelli，又名滑翔树蛙）的卵团数据。该湿地主体区域（坐标：08° 28.530’ N, 083° 35.566’ W）面积约为45×20米，另有一片狭窄浅滩区域向西延伸约20米。两个叶泡蛙属物种均因强降雨启动繁殖。本次采样持续6天：红眼树蛙的最早孵化时间为4天，两个物种的孵化期普遍为5~8天。

野外采集的数据包括产卵时间、产卵位点与捕食相关信息（卵团的整体发育状态、大小，产卵位点的基质、朝向、相对于冠层与水面的位置，以及捕食证据等）。我们在湿地内及周边的植物上搜寻卵团，搜寻高度范围为水面至约3米高处。由于无法可靠观测冠层更高处的卵团，且已有文献表明滑翔树蛙的绝大多数卵团均位于3米以下高度（Scott与Starrett，1974），因此本次发现的所有卵团均处于距水面1.6米以内的区域。我们对每个卵团的位置进行临时标记，并每日到访以统计胚胎数量，同时记录捕食、病害或发育异常的相关证据。当出现直接捕食证据（胚胎与卵胶膜缺失、卵破裂），或发育早期（基于Cohen等人2016年的研究，即孵化前4天内）单日胚胎数量变化超过10枚时，我们将其判定为发生捕食事件。我们将捕食类型划分为无捕食、卵团体积缺失（疑似蛇类捕食）、卵胶膜留存但胚胎缺失（若发生在发育早期）以及卵破裂（疑似无脊椎动物捕食）。对于发育4天后仍留存卵胶膜但胚胎缺失的卵团，我们判定其已开始自然孵化。

我们根据卵团被发现时的发育状态，计算了每个卵团的额外变量，包括体积、胚胎密度与产卵日期。我们基于椭球体公式计算卵团体积：体积= (4/3×π×a×b×c)/2，其中a、b、c分别为卵团的长、宽、高半径。对于产在叶片或相对平坦表面上的卵团，取半椭球体体积；对于包裹在细枝或藤蔓上的卵团，则取完整椭球体体积。仅针对初始计数时未出现卵胶膜缺失等捕食迹象的卵团，我们计算了其胚胎密度（统计的最大胚胎数/卵团体积）。对于未被即时发现的卵团，我们通过将孵化日期减去7（两个物种的平均孵化时间，Savage，2002）来估算其产卵日期。采样最后一天的前一晚发生了强降雨，红眼树蛙随即出现新一轮产卵。为避免高估统计检验的置信度，我们在产卵日期分析中未纳入最后一天产出的卵团。

本研究经德克萨斯理工大学动物护理与使用委员会批准（审批号：17063-07），并获得针对A. Vega的研究许可INV-ACOSA-053-17，所有操作均符合相关机构与国家的实验规范。