Predation risk estimated on live and artificial insect prey follows different patterns

Models mimicking prey organisms are increasingly used in ecological studies including testing fundamental ecological and evolutionary theories. The general consensus is that predation risk estimated on artificial models may not quantitatively correspond to predation pressure on live prey, but it still can be used in various comparisons. We tested whether the use of live and artificial prey reveals the same patterns of variation in predation risk. We exposed live prey (blowfly larvae and puparia) and plasticine models of blowfly puparia in two boreal forest sites, both openly and in ant- and bird-exclusion treatments, and we quantified attacks by both avian and invertebrate predators. Bird attack rates were always higher on live puparia than on their plasticine models, but the magnitude of this difference declined from 8.4-fold in early summer to 2-fold in mid- and late summer. We attribute these changes to different responses to prey by experienced adult birds that dominate the bird communities in early summer versus explorative juvenile birds that are abundant later in the season. Invertebrate daily predation rates on maggots decreased from 56% in early summer to 28% in late summer, but invertebrate attacks on plasticine models showed no seasonal changes. Overall, invertebrate predation on maggots was 67-fold greater than their predation on models. Observations showed that wood ants did not attack plasticine models and did not leave on them any damage marks. Estimates based on artificial prey indicate a much greater role of bird predation than invertebrate predation, while estimates based on live prey suggest the opposite pattern. Thus, using live and artificial prey may lead to different conclusions about relative importance of different predator groups in a locality. Moreover, for both avian and invertebrate predators, predation risk based on artificial and live prey shows different seasonal changes and may potentially demonstrate different spatial patterns.

Methods
In our study, we used live larvae (maggots hereafter) and puparia of the blowfly Calliphora sp. (Diptera: Calliphoridae) and plasticine models with similar appearance to fly puparia in size, colour and shape. Maggots, obtained in a local fishing bait shop, were 10–13 mm long. They were divided into two equal parts: one part was placed in a refrigerator (at 4°C) to prevent pupation and the other part was kept at room temperature (22°C) and allowed to pupate. We mixed red and black plasticine (advertised as non-toxic and odourless; Chemical plant “Luch”, Yaroslavl, Russia) in the proportion 4:1 to obtain a colour (brown) similar to that of puparia, and we moulded models from this plasticine. Three individual models were attached in a line along a stick (about 10 cm long, cut from thin birch branches) using universal glue (Yleisliima; Biltema Suomi Oy, Helsinki, Finland). In the same fashion we glued three puparia. The live and plasticine puparia set in this way mimicked some caterpillar-like prey.

The data were obtained from three different experiments. Experiment 1 was designed to compare bird and invertebrate predation on three different prey items. For this experiment, at each site, we selected 10 pairs (blocks) of young (2–3 m tall) downy birches (Betula pubescens). The distance between the paired trees was 0.5–2 m, and the distance between blocks was about 20 m. One tree in each block was isolated from non-flying invertebrate predators (mostly ants) by applying a ring of non-drying glue (Sticky-Trap, Vilofarm A/S, Hobro, Denmark) at the base of the trunk (ant exclusion hereafter). All ants and other invertebrates were removed by shaking these isolated trees and then manually. We ensured that the branches and foliage of any ant-exclusion plants did not touch other trees to prevent any migration of non-flying invertebrate predators (with possible exception of spiders) from neighbouring trees. The second tree in each block remained accessible to ants and other predators (control hereafter).

On each of 20 trees, we placed three kinds of prey items: an open transparent plastic vial (40 mm height and 30 mm diameter) containing three maggots, one stick with three puparia and one stick with three models. The vial was attached with tape to the tree trunk 80–100 cm above the ground. Each vial also contained a small stick to ease the movement of predatory invertebrates (we had previously verified that the maggots were incapable of climbing on the stick). Sticks with puparia and models were attached with small pieces of wire to birch branches at 100–150 cm above the ground at the outer part of the crown, so that the distance between them was as far as possible. Experiment 1 was repeated six times in Kustavi and seven times in Turku using the same trees from 28 May to 11 September 2021. At 24 h from the start of the experiment, we recorded (i) the number of maggots that had disappeared from each vial, (ii) the number of dead maggots in each vial; (iii) the number of damaged puparia; and (iv) the presence of attack marks by both invertebrates and birds on each of the three plasticine models. At 4–7 days from the start of the experiment, we repeated records on live puparia and plasticine models.

Experiment 2 was conducted to explore whether birds attacked the maggots exposed in vials. At the beginning of June 2021, at both study sites, we selected 20 young birches and attached a vial with five maggots to the trunk, as described above. To exclude bird predation, we covered the openings of the vials on 10 of these birches with bird netting. The number of maggots that disappeared from vials was recorded at 2 days from the start of the experiment.

Experiment 3 was conducted to compare invertebrate attacks on maggots and puparia. In early June of 2021, we placed two vials (as in experiment 1) on each of 20 birch trees in each of two sites. Half of the trees were isolated from ants by glue rings, as described above. We placed three maggots in one of these vials and three puparia into the other. The number of prey individuals that disappeared from vials was recorded at 4 days from the start of the experiment.

模拟猎物生物的模型正日益广泛应用于生态学研究，其中包括验证基础生态学与进化生物学理论。学界普遍认为，基于人工模型估算的捕食风险，虽无法在定量层面对应活体猎物所承受的捕食压力，但仍可用于各类对比研究。本研究旨在验证：使用活体与人工猎物是否能得出一致的捕食风险变异格局。

我们在两处北方森林样地中布设了活体猎物（丽蝇（blowfly）幼虫与蛹）以及丽蝇蛹的橡皮泥模型，同时设置了开放式、排除蚂蚁和鸟类的处理组，并量化了鸟类与无脊椎动物捕食者的攻击事件。结果显示，鸟类对活体蛹的攻击率始终高于其橡皮泥模型，但该差异的倍数从初夏的8.4倍降至仲夏与晚夏的2倍。我们将这一变化归因于：初夏占据鸟类群落主导地位的经验丰富的成鸟，与季末数量占优的探索性亚成鸟，对猎物的响应存在差异。无脊椎动物对蛆虫的日均捕食率从初夏的56%降至晚夏的28%，但无脊椎动物对橡皮泥模型的攻击未表现出季节性变化。总体而言，无脊椎动物对蛆虫的捕食强度是其对模型的67倍。观察显示，林地蚂蚁不会攻击橡皮泥模型，也不会在其表面留下任何损伤痕迹。基于人工猎物的估算结果显示，鸟类捕食的作用远大于无脊椎动物捕食；而基于活体猎物的估算则呈现相反的格局。因此，在某一区域内，使用活体与人工猎物可能会对不同捕食类群的相对重要性得出截然不同的结论。此外，对于鸟类与无脊椎动物捕食者而言，基于人工与活体猎物的捕食风险不仅呈现出不同的季节性变化，还可能表现出不同的空间格局。

方法
本研究使用了丽蝇属（Calliphora sp.，双翅目：丽蝇科）的活体幼虫（下文简称蛆虫）与蛹，以及尺寸、颜色、形态均与蝇蛹相似的橡皮泥模型。从当地渔具店购得的蛆虫体长为10–13 mm。将其分为均等的两部分：一部分置于4℃冰箱中以阻止化蛹，另一部分置于室温（22℃）环境下使其完成化蛹。我们按照4:1的比例混合红色与黑色橡皮泥（宣称无毒无异味；俄罗斯雅罗斯拉夫尔卢赫化工厂生产），调配出与蝇蛹相近的棕褐色，并以此塑形为模型。使用万能胶（Yleisliima；芬兰赫尔辛基Biltema Suomi Oy公司生产），将3个独立模型沿一根约10 cm长的细桦树枝排成一条直线固定。我们以相同方式固定3个活体蛹。以此方式布设的活体蛹与橡皮泥蛹，可模拟部分类似毛虫的猎物。

本研究的数据来自3组独立实验。
实验1旨在对比3类不同猎物的鸟类与无脊椎动物捕食情况。针对该实验，我们在每个样地中选取10组（区组）高度为2–3 m的毛桦（Betula pubescens），每组内两树间距为0.5–2 m，区组间间距约为20 m。每个区组中的1株树会在树干基部涂抹一圈不干胶（Sticky-Trap；丹麦霍布罗Vilofarm A/S公司生产），以阻隔非飞行性无脊椎动物捕食者（主要为蚂蚁），即后文所述的蚂蚁排除组。通过摇晃并辅以人工清理的方式，移除该隔离树上的所有蚂蚁与其他无脊椎动物。我们确保蚂蚁排除组植株的枝条与叶片不会接触其他树木，以防止非飞行性无脊椎动物捕食者（蜘蛛可能除外）从邻树迁移至此。每个区组中的另一株树则保持对蚂蚁与其他捕食者开放，即后文所述的对照组。

我们在20株树上分别布设3类猎物：装有3只蛆虫的开放式透明塑料瓶（高40 mm，直径30 mm）、带有3个蛹的桦树枝条，以及带有3个模型的桦树枝条。该塑料瓶用胶带固定于树干离地80–100 cm处。每个塑料瓶内还放置了一根小木棍，以方便捕食性无脊椎动物活动（我们此前已验证蛆虫无法攀爬该木棍）。带有蛹或模型的桦树枝条则用细铁丝固定于树冠外围离地100–150 cm的枝条上，尽可能拉开彼此间距。实验1于2021年5月28日至9月11日期间，在库斯塔维（Kustavi）重复开展6次，在图尔库（Turku）重复开展7次，使用的均为同一批树木。实验开始后24小时，我们记录以下内容：(i) 每个塑料瓶中消失的蛆虫数量；(ii) 每个塑料瓶内死亡的蛆虫数量；(iii) 受损的活体蛹数量；(iv) 3个橡皮泥模型上是否存在无脊椎动物与鸟类的攻击痕迹。实验开始后4–7天，我们再次记录活体蛹与橡皮泥模型的相关数据。

实验2旨在探究鸟类是否会攻击塑料瓶内暴露的蛆虫。2021年6月初，我们在两个研究样地各选取20株幼桦树，按照前述方式将装有5只蛆虫的塑料瓶固定于树干上。为排除鸟类捕食，我们用鸟网覆盖其中10株树所绑定塑料瓶的瓶口。实验开始后2天，记录塑料瓶中消失的蛆虫数量。

实验3旨在对比无脊椎动物对蛆虫与蛹的攻击情况。2021年6月初，我们在两个样地各选取20株桦树，每株放置2个与实验1一致的塑料瓶。其中一半树木按照前述方式用胶圈隔离蚂蚁。我们在其中一个塑料瓶中放入3只蛆虫，另一个塑料瓶中放入3个蛹。实验开始后4天，记录塑料瓶中消失的猎物个体数量。