Birds host a vast diversity of feather symbionts of different kingdoms, including animals (e.g., lice, mites), fungi, and bacteria. Feather mites (Acariformes: Astigmata: Analgoidea and Pterolichoidea), the most abundant animal ectosymbionts of birds, are permanent inhabitants of the pterosphere (ptero feather in Greek; Labrador et al. 2020), and the ones studied here are easily spotted as small (ca. 0.5 mm) dots on the surface of flight feathers. They are highly host specific symbionts (Doña et al. 2018), and they seem to be commensals or even mutualists of birds by taking detritus and microorganisms such as fungi and bacteria from feathers, some of which are keratinophilic and therefore can damage the feathers (Blanco and Tella 1997, Galván 2012, Doña et al. 2019). However, many basic questions remain to be answered, such as the moments and the places where feather mites eat. Indeed, we wondered whether this might be partly because feather mites have been studied mainly during the day, when (most) birds fly, rather than during the night when mites seem to move more freely on the wings, according to two old anecdotal reports (Dubinin 1951, McClure 1989). To investigate the night ecology of feather mites, we initially spent a whole night observing them on two individual birds. At that point, we were unaware of how it would change our understanding of the pterosphere.

On February 13, 2020, we captured two blackcaps Sylvia atricapilla (Linnaeus, 1758) (Sylviidae), with mist nets at 17 h (still in daylight). To quantify the mites and study their spatial ecology, we held the extended wings and tail of the birds in front of a white paper with a light source behind, and took photographs of each primary (10 feathers), secondary (6), and tertial (3) feather of the two wings, and the 12 tail feathers, placing a ruler close to the feathers for subsequent measurements. Then, we kept the birds outdoors overnight (minimum night temperature 4°C) in profusely perforated paperboard boxes, repeating the photographs at 20 h (night already), 23 h, 2 h, 5 h, and 9 h (daylight again) (see Appendix S1: Section S1 for methodological details). When we saw the abundance and unusual (for us) distribution of mites on the first blackcap at 20 h, we guessed that many findings were ahead, and not merely about their nocturnal occurrence on the wings.

By studying the abundance of mites, we found that (Fig. 1a): (1) The total number of mites increased on the wings and tail until midnight and then decreased toward dawn to reach abundances on the wings that were even lower than those recorded at dusk. (2) Different mite species and developmental stages showed different abundance dynamics: adult Proctophyllodes sylviae Gaud, 1957 (Proctophyllodidae) initially slightly increased in numbers, decreasing afterward. Similarly, juvenile P. sylviae suddenly appeared at night on the wings and tail and almost disappeared again at dawn. Lastly, Trouessartia sp. (likely T. bifurcata (Trouessart, 1885) (Trouessartiidae) based on host records in Doña et al. 2016) did not show the initial increase found in P. sylviae, but instead gradually decreased during the night. These results supported the previous non-quantitative records of mites increasing at dusk and moving along the wings overnight (Dubinin 1951, McClure 1989), and show that different species and stages behave differently during the nighttime.

To further analyze the nocturnal activity of feather mites, we digitized their position on the feathers from the photographs (totaling 4,643 mite coordinates). We discovered: (1) Mites migrated toward the feather tips (which were seldom occupied at 17 h, and are not usually occupied by blackcap feather mites during the day; ML, JD, DS, RJ, personal observation). This pattern was mainly observed in primaries (but also in secondaries) and was followed by a progressive return to more proximal positions toward dawn (Fig. 1b; Appendix S1: Fig. S1). (2) Mites tightly aggregated by developmental stage and species among and within feathers (Appendix S1: Fig. S1). This distribution was especially apparent in P. sylviae juveniles, which were clustered in certain feathers and particular locations within feathers (Appendix S1: Fig. S1). (3) Mites at night occupied central tail feathers, which are rarely occupied during the day (ML, JD, DS, RJ, personal observation, Appendix S1: Fig. S1). (4) During the night, P. sylviae occurred on the ventral surface of feathers and Trouessartia sp. occurred on the dorsal one, as they did during the daytime and as previously found (Dabert and Mironov 1999, Proctor 2003, Mestre et al. 2011, Fernández-González et al. 2013; ML, JD, DS, RJ, personal observation).

The nocturnal presence and activity of feather mites on wing and tail feathers made us wonder whether mites could be feeding during the night. To test this, on February 20, 2020, we sampled feather mites at dusk and then at dawn from two blackbirds Turdus merula Linnaeus, 1758 (Turdidae) and another blackcap, and investigated mite gut contents under the microscope (see Appendix S1: Section S2 for methodological details). Many mites had no visible food items in their esophagus and ventriculus (first gut compartments where food arrives when ingested; Alberti and Dabert 2012), but many other mites had their esophagus and ventriculus fully packed with food (Fig. 2a). To test whether mites taken at dawn were those full of food (i.e., which would show night feeding), we took pictures from 120 randomly selected adult female Proctophyllodes mites (P. musicus Vitzthum, 1992 from blackbirds; P. sylviae from blackcaps), 20 from each of the three birds and the two sampling times. Then, we coded the slides so that three of us could blindly sort the pictures according to the amount of food found in the esophagus and ventriculus. Afterward, we discussed our minor discrepancies to achieve a consensus rank among mites and counted the number of food boluses and fecal pellets as a measure of previous feeding events (Alberti and Dabert 2012).

We found an almost perfect division between the mites sampled at dusk and those at dawn (Fig. 1c vertical colored lines, Fig. 2a). At dusk, mites had either no food items (31.7%) or <1/3 of the ventriculus with food items (56.7%). In contrast, at dawn, most mites had 1/3–2/3 (28.3%) or >2/3 (60.0%) of the ventriculus full of food items (Figs. 1c, 2a). Moreover, this pattern was consistent across the three individual birds studied (Fig. 1c vertical colored lines). Conversely, most mites sampled at dusk (76.7%) and dawn (91.7%) had one food bolus and one fecal pellet (Fig. 2a; Appendix S1: Fig. S2).

Later, when working on Fig. 2a, we noticed a large but tenuous and transparent egg inside one of the females (Fig. 2b). We went through the 120 pictures again and then discovered 21 females with an egg inside. Perhaps more interestingly, while as many as 20 out of the 60 (33.3%) females collected at dusk carried an egg, only one out of 60 (1.7%) did so at dawn. Again, this pattern was consistent in the three birds sampled: the blackcap (dusk: 5 females with egg/15 without egg, dawn: 1/19), the first blackbird (dusk: 7/13, dawn: 0/20), and the second blackbird (dusk: 8/12, dawn: 0/20).

The results presented here are based on a large sample of feather mites but a low number of individual birds and of mite and bird species, and will therefore need further exploration in dedicated studies that include a larger sample of birds (and therefore also of mites) of different species. However, the nocturnal patterns of abundance, spatial distribution, feeding behavior, and egg-laying that we have reported here suggest that feather mites are nocturnal organisms. This has several implications:

First, it could be that survey of feather mite abundance on individual birds are systematically underestimated because the most common method of assessment consists of counting mites on wing and tail feathers during the daytime. This underestimation may potentially differ between mite species and stages, as Fig. 1a suggests. Importantly, further studies are needed to ascertain whether this day/night difference in feather mite counts varies among bird species. If so, current data on feather mite abundances across bird species may be potentially biased (e.g., Galván et al. 2012, Diaz-Real et al. 2014).

Second, we have found that mites were able to move from the center of primary feathers to more distal positions in three hours (Appendix S1: Fig. S1), confirming that they have the potential to change their position rapidly (e.g., in response to changes in environmental conditions, Dubinin 1951, McClure 1989, Wiles et al. 2000). Furthermore, feather mites presented a non-random nocturnal spatial distribution among and within feathers (Appendix S1: Fig. S1) and showed clear location preferences (as they do during the day, e.g., Pérez and Atyeo 1984), which also differed between mite species and stages. Thermo-orientation could potentially be a driver of these location changes, as has been found in wing feather lice. Specifically, some lice have been seen to rapidly change position in search of specific feather regions with more suitable temperatures (that vary depending on lice life stage) (Harbison and Boughton 2014).

The migration of mites toward the tip of the primary feathers in the first half of the night is intriguing. A potential explanation would be that feather tips accumulate more organic particles to feed upon. Another possibility is that mites would take up water by moving to feather tips (dangerous during the day in a flying bird), as these feather regions are more exposed to air humidity when the wing is folded in a roosting bird. Indeed, Gaede and Knülle (1987) found atmospheric water vapor to be the most important water source for Proctophyllodes troncatus Robin, 1877 (the only feather mite species in which this aspect has been studied), as also occurs in bird lice (Rudolph 1983).

Our observations of food items in mites under the microscope also have important implications (Figs. 1c, 2a; Appendix S1: Fig. S2). In short, feather mites collected at dusk had empty esophagus–ventriculus, one food bolus, and a fecal pellet. The mites sampled at dawn had full esophagus–ventriculus and (again) one food bolus and one fecal pellet. Therefore, it is reasonable to suggest that feather mites feed at night, and during the day the food in the esophagus–ventriculus becomes a food bolus, the food bolus becomes a fecal pellet, and one fecal pellet is egested. If this proposed process was correct, it would have several implications:

First, most mites feed every night, given that almost all mites had a medium–high abundance of food items in the esophagus–ventriculus at dawn (Fig. 1c), and one food bolus and a fecal pellet (see above; Appendix S1: Fig. S2). This potential need for filling the ventriculus every night poses the scenario of feather mites being sensitive to food shortages, and therefore population dynamics of feather mites being (at least partly) bottom-up regulated by food resources found on feathers.

Second, by taking advantage of our data, we can make a very rough estimation of the amount of organic material that feather mites could clean from feathers' surface (see Appendix S1: Section S3 for the calculus details). Every night, feather mites could eliminate an average of ca. 0.17 mm² of compacted (as found in a vacuum cleaner filter) fungi, bacteria, and other organic particles from the wing flight feathers of a single blackbird. This would lead to ca. 0.62 cm² every year on a blackbird and a rough yearly estimate of ca. 80,000 m² in just European passerines. Most importantly, these back-of-the-envelope calculations showed that while bacteria, fungi, and feather mites are microscopic, feather mites can daily clean feathers at a macroscopic scale. This result reinforces the idea that feather mites play a role in cleaning bird feathers (Doña et al. 2019), and that they may top-down regulate pterosphere's microorganisms.

While preliminary, our findings suggest that a major adaptation for living on bird flight feathers is to undertake important biological functions such as feeding or egg-laying during the night (i.e., when it is safer to move along flight feathers). This should encourage nocturnal studies on feather mites and other bird ectosymbionts, which often show convergent adaptations (Jovani 2003). Indeed, pigeon wing lice also seem to change their abundance on wing feathers at night (Sarah E. Bush, personal communication), encouraging further studies on the selective value of this behavior. Overall, our exploration of the nocturnal natural history of these few species of feather mites may trigger a major shift in our understanding of the pterosphere.

鸟类拥有来自不同王国的多种多样的羽毛共生体，包括动物（例如虱子、螨虫）、真菌和细菌。羽螨（Acariformes: Astigmata: Analgoidea and Pterolichoidea）是鸟类最丰富的动物外共生体，是翼状圈的永久居民（希腊语为ptero feather；Labrador et al. 2020），而这里研究的那些很容易被发现很小（约 0.5 mm) 飞羽表面的点。它们是高度宿主特异性的共生体（Doña et al. 2018），它们似乎是鸟类的共生体，甚至是鸟类的共生体，通过从羽毛中获取碎屑和微生物（如真菌和细菌），其中一些是角质素的，因此会损坏羽毛（布兰科和泰拉1997，Galván 2012 年，Doña 等人。2019 年）。然而，许多基本问题仍有待回答，例如羽毛螨进食的时刻和地点。事实上，我们想知道这是否可能部分是因为羽毛螨主要在白天进行研究，当时（大多数）鸟类飞行，而不是在夜间螨似乎更自由地在翅膀上移动，根据两份古老的轶事报告（杜比宁1951 年，麦克卢尔1989 年）。为了调查羽毛螨的夜间生态，我们最初花了一整夜观察它们在两只单独的鸟身上。那时，我们还没有意识到它会如何改变我们对翼状圈的理解。

2020 年 2 月 13 日，我们抓获了两只黑帽Sylvia atricapilla(Linnaeus, 1758) (Sylviidae)，17 小时有雾网（仍在白天）。为了量化螨虫并研究它们的空间生态学，我们将鸟类伸展的翅膀和尾巴放在一张白纸前，后面有一个光源，并拍摄了每根主要（10 根羽毛）、次要（6 根）和三次羽毛的照片(3) 两个翅膀的羽毛和 12 根尾羽，将尺子放在靠近羽毛的位置进行后续测量。然后，我们将鸟放在带孔的纸板箱中过夜（夜间最低温度 4°C），在 20 小时（已经是夜晚）、23 小时、2 小时、5 小时和 9 小时（又是白天）重复拍摄照片（有关方法的详细信息，请参见附录 S1：第 S1 节）。当我们在 20 小时看到第一个黑帽上螨虫的丰富和不寻常的（对我们而言）分布时，我们猜测许多发现都在前面，

通过研究螨虫的丰度，我们发现（图1a）：（1）直到午夜，翅膀和尾巴上的螨虫总数增加，然后到黎明时减少，达到翅膀上的丰度，甚至低于记录的数量傍晚。(2)不同的螨种和发育阶段表现出不同的丰度动态：成虫Proctophyllodes sylviae Gaud, 1957(Proctophyllodidae)的数量最初略有增加，随后逐渐减少。同样， 少年P。sylviae在晚上突然出现在翅膀和尾巴上，在黎明时分几乎又消失了。最后，Trouessartia sp。（可能是T .  bifurcata(Trouessart, 1885) (Trouessartiidae) 基于 Doña 等人的宿主记录。2016 年）没有显示P的初始增长。 sylviae，而是在夜间逐渐减少。这些结果支持了之前关于螨虫在黄昏时增加并在夜间沿着翅膀移动的非定量记录（Dubinin 1951，McClure 1989），并表明不同物种和阶段在夜间表现不同。

为了进一步分析羽毛螨的夜间活动，我们将照片中它们在羽毛上的位置数字化（总共 4,643 个螨坐标）。我们发现：（1）螨虫向羽尖迁移（17h很少被占据，白天通常不被黑头羽螨占据；ML，JD，DS，RJ，个人观察）。这种模式主要在初级（但也在次级）中观察到，随后逐渐返回到接近黎明的位置（图 1b；附录 S1：图 S1）。(2) 羽毛间和羽毛内的发育阶段和物种紧密聚集的螨虫（附录S1：图S1）。这种分布在P中尤为明显。 西尔维娅幼鸟，它们聚集在某些羽毛和羽毛内的特定位置（附录S1：图S1）。（3）螨虫夜间占据中央尾羽，白天很少占据（ML，JD，DS，RJ，个人观察，附录S1：图S1）。(4) 夜间，P. sylviae出现在羽毛和Trouessartia sp.的腹面。发生在背侧，就像他们在白天所做的那样，正如之前发现的那样（Dabert and Mironov 1999 , Proctor 2003 , Mestre et al. 2011 , Fernández-González et al. 2013 ; ML, JD, DS, RJ,个人观察） .

羽螨在翼羽和尾羽上的夜间存在和活动使我们想知道螨虫是否可以在夜间觅食。为了验证这一点，我们在 2020 年 2 月 20 日从两只黑鹂Turdus merula Linnaeus、1758（Turdidae）和另一只黑鹂中分别在黄昏和黎明时采集了羽螨样本，并在显微镜下研究了螨虫的肠道内容物（见附录 S1：S2 部分）方法细节）。许多螨虫的食道和脑室没有可见的食物（食物摄入时到达的第一个肠道隔室；Alberti 和 Dabert 2012），但许多其他螨虫的食道和脑室都充满了食物（图2a）。为了测试在黎明时拍摄的螨虫是否是那些充满食物的螨虫（即，它们会显示夜间进食），我们从 120 只随机选择的成年雌性Proctophyllodes螨虫（P. musicus Vitzthum ,  1992 from blackbirds; P.  sylviae ）中拍摄了照片来自黑帽），来自三只鸟中的每只和两个采样时间的 20 个。然后，我们对幻灯片进行编码，以便我们三个人可以根据食道和脑室中的食物量对图片进行盲目排序。之后，我们讨论了我们的微小差异，以在螨虫中达成共识等级，并计算食物团和粪便颗粒的数量作为先前喂养事件的衡量标准（Alberti 和 Dabert 2012）。

我们发现在黄昏和黎明时采样的螨虫之间几乎完美的划分（图 1c 垂直彩色线，图 2a）。黄昏时，螨虫要么没有食物（31.7%），要么只有不到 1/3 的脑室有食物（56.7%）。相比之下，在黎明时分，大多数螨虫有 1/3-2/3 (28.3%) 或 >2/3 (60.0%) 的脑室充满食物（图 1c、2a）。此外，这种模式在所研究的三只单独的鸟类中是一致的（图 1c 垂直彩色线）。相反，在黄昏（76.7%）和黎明（91.7%）采样的大多数螨虫有一个食物团和一个粪便颗粒（图2a；附录S1：图S2）。

后来，在处理图 2a 时，我们注意到其中一个雌性体内有一个大而脆弱且透明的卵（图 2b）。我们再次浏览了这 120 张照片，然后发现了 21 名女性，里面有一个鸡蛋。或许更有趣的是，在黄昏收集的 60 只雌性中，多达 20 只（33.3%）携带了一个卵，而在黎明时分，60 只（1.7%）只携带了一个卵。同样，这种模式在采样的三只鸟中是一致的：黑帽（黄昏：5 只雌性有蛋/15 只没有蛋，黎明：1/19），第一只黑鹂（黄昏：7/13，黎明：0/20），和第二只黑鸟（黄昏：8/12，黎明：0/20）。

这里提供的结果基于大量羽毛螨样本，但个体鸟类以及螨虫和鸟类物种数量较少，因此需要在包括更大鸟类样本（因此也包括螨虫）的专门研究中进一步探索不同种类的。然而，我们在此报道的丰度、空间分布、摄食行为和产卵的夜间模式表明羽毛螨是夜间生物。这有几个含义：

首先，可能系统地低估了对个体鸟类羽毛螨丰度的调查，因为最常见的评估方法包括在白天计算翼羽和尾羽上的螨虫。如图1a所示，这种低估可能在螨种和阶段之间存在差异。重要的是，需要进一步研究以确定羽毛螨数量的昼夜差异是否因鸟类而异。如果是这样，当前有关鸟类羽毛螨丰度的数据可能存在偏差（例如，Galván 等人2012 年，Diaz-Real 等人2014 年）。

其次，我们发现螨虫能够在三个小时内从初级羽毛的中心移动到更远的位置（附录 S1：图 S1），证实它们有可能快速改变位置（例如，响应环境条件的变化，Dubinin 1951 , McClure 1989 , Wiles et al. 2000 )。此外，羽毛螨在羽毛之间和羽毛内呈现出非随机的夜间空间分布（附录 S1：图 S1）并显示出明确的位置偏好（就像它们在白天所做的那样，例如，Pérez 和 Atyeo 1984)，这在螨种和阶段之间也有所不同。正如在翼羽虱中发现的那样，热定向可能是这些位置变化的驱动因素。具体来说，一些虱子会迅速改变位置以寻找具有更合适温度的特定羽毛区域（取决于虱子的生命阶段）（Harbison 和 Boughton 2014）。

螨虫在前半夜向初级羽毛尖端的迁移是耐人寻味的。一种可能的解释是，羽毛尖端会积累更多的有机颗粒以供食用。另一种可能性是螨虫会通过移动到羽毛尖端来吸收水分（在白天对飞鸟来说很危险），因为当翅膀折叠在栖息的鸟类中时，这些羽毛区域更容易暴露在空气湿度中。事实上，Gaede 和 Knülle ( 1987 ) 发现大气中的水蒸气是Proctophyllodes troncatus Robin, 1877（唯一研究过这方面的羽螨物种）最重要的水源，鸟虱中也存在这种情况（Rudolph 1983） .

我们在显微镜下对螨虫食物的观察也具有重要意义（图 1c、2a；附录 S1：图 S2）。简而言之，黄昏时收集的羽螨有空的食道-心室、一个食物团和一个粪便颗粒。黎明时取样的螨虫有完整的食道 - 心室和（再次）一个食物团和一个粪便颗粒。因此，有理由建议羽螨在夜间进食，白天食道-心室中的食物变成食物团，食物团变成粪便颗粒，吞下一个粪便颗粒。如果这个提议的过程是正确的，它将有几个含义：

首先，大多数螨虫每晚都进食，因为几乎所有螨虫在黎明时食道-心室中都有中高丰度的食物（图 1c），以及一个食物团和一个粪便颗粒（见上文；附录 S1：图 S2)。这种每晚填充脑室的潜在需求造成了羽毛螨对食物短缺敏感的情况，因此羽毛螨的种群动态（至少部分）由羽毛上的食物资源自下而上调节。

其次，利用我们的数据，我们可以非常粗略地估计羽毛螨可以从羽毛表面清除的有机物质的数量（有关微积分的详细信息，请参见附录 S1：S3 部分）。每天晚上，羽毛螨平均可以消灭大约 10 公斤。0.17 mm ²压实的（如在真空吸尘器过滤器中发现的）真菌、细菌和其他有机颗粒，来自一只黑鸟的翅膀羽毛。这将导致约。黑鹂每年0.62 cm ²和大约每年的粗略估计。80,000 米²在只是欧洲雀形目。最重要的是，这些粗略的计算表明，虽然细菌、真菌和羽毛螨是微观的，但羽毛螨可以在宏观范围内每天清洁羽毛。这一结果强化了羽毛螨在清洁鸟类羽毛中发挥作用的观点（Doña 等人，2019 年），并且它们可能自上而下地调节翼球圈的微生物。

虽然是初步的，但我们的研究结果表明，靠鸟类飞行羽毛生活的一个主要适应是承担重要的生物学功能，例如在夜间进食或产卵（即，当沿着飞行羽毛移动更安全时）。这应该鼓励对羽毛螨和其他鸟类外共生生物进行夜间研究，这些研究通常表现出趋同的适应（Jovani 2003）。事实上，鸽翅虱似乎也在夜间改变了它们在翅膀羽毛上的数量（Sarah E. Bush，个人交流），鼓励进一步研究这种行为的选择性价值。总体而言，我们对这几种羽毛螨的夜间自然历史的探索可能会引发我们对翼状圈理解的重大转变。