How will anthropogenic changes (species invasions/extinctions, land‐use conversion, and climate change) influence the pollination and reproductive success of the world's angiosperms, 85% of which require animal pollination (Ollerton et al., 2011)? Answering this question requires understanding the mechanisms that cause pollen limitation of seed production. Pollen limitation (PL) is widespread (Bennett et al., 2018) and occurs when pollinators fail to deliver adequate quantity or quality of pollen to stigmas (Ashman et al., 2004). It is thought to primarily occur when there are few pollinators visiting the plants or when the pollinators that visit do not bring enough conspecific pollen (CP) or bring CP that is of low quality (e.g., self pollen that comes from a different flowers on the same plant) (Ashman et al., 2004; Aizen and Harder, 2007). Often left out of the conversation on PL, however, is that pollinators can also transfer heterospecific pollen (HP) among species (but see Jakobsson et al., 2009; McKinney and Goodell, 2010).

Studies of HP receipt on stigmas collected late in anthesis demonstrate that almost all plants (88%) receive at least some HP, and when they do, it can make up to 75% of the total stigmatic pollen load (Ashman and Arceo‐Gómez, 2013; Arceo‐Gómez et al., 2019). The HP deposited on stigmas can interfere with CP tube growth and reduce CP germination and fertilization (Morales and Traveset, 2008), though the effects can vary widely (Ashman and Arceo‐Gómez, 2013). Thus, HP can be viewed as the lowest quality pollen that a flower might receive as it cannot result in legitimate seed production and can even reduce seed set. Moreover, if there is a trade off between HP and CP transport on pollinators’ bodies, greater HP transfer can lead to deposition of fewer CP on a conspecific stigma (e.g., Moreira‐Hernández and Muchhala, 2018). Anthropogenic environmental changes may influence PL through changes in the quantity and quality of CP, as well as the quantity (and identity) of HP delivered by pollinators. Predicting which plant species will be most vulnerable to PL in the future requires an understanding of both these mechanisms.

Traditional means to measure PL involve randomly assigning flowers to open, naturally pollinated (control) and hand‐pollinated (supplemental) treatments, and subsequently measuring the degree to which reproductive success is increased in the supplement treatment (Bennett et al., 2018). Because the pollen supplementation treatment is applied to stigmas that are also open to natural pollination (Ashman et al., 2004; Bennett et al., 2018), these stigmas may have previously received or may subsequently receive both CP and HP. There are several ways that HP could affect estimates of PL. First, higher reproductive success in the supplement treatment, which indicates PL, could be due not to insufficient CP but rather due to strong HP interference in the open flowers (not supplemented); the CP supplement treatment increases the ratio of CP to HP and thus decreases the importance of HP interference (Fig. 1A). Another possibility is that the higher reproductive success in the supplement treatment could be due to the supplemental CP (in recently opened flowers) preventing subsequent HP deposition/adherence, which may be common in open flowers that are not supplemented with CP (Fig. 1B). Last, some researchers measure PL by comparing flowers of open‐pollinated plants to those that are hand‐outcrossed and then bagged to prevent pollinator visitation (Bennett et al., 2018; Fig. 1C). In the extreme case in these experiments, the open treatment might have both HP and low‐quality CP compared to the supplement treatment, which only receives the high‐quality CP given in the treatment (the bagging prevents anything else from arriving on the stigma). It is worth noting that even these scenarios oversimplify the dynamics of natural pollen transfer. Nevertheless, they still reveal the potential complexity of interspecific pollen interactions on the stigma (between CP and HP) that result from pollen transport on pollinator bodies (Arceo‐Gómez and Ashman, 2011; Minnaar et al., 2018) and that can impact PL.

A common thought is that plants evolve to reduce PL and that PL is evidence of a contemporary disruption from their evolutionarily stable optima allocation strategy, perhaps owing to loss of pollinators and reduced delivery of high‐quality CP (reviewed by Ashman et al., 2004). Indeed, anthropogenic changes can cause severe PL, especially for pollinator‐dependent species (Burns et al., 2019; J. M. Bennett et al., unpublished manuscript). However, we do not know the degree to which the evidence of PL due to anthropogenic change reflects reduced CP quantity, reduced CP quality, and/or increased HP receipt. An indication that increased HP receipt might play an important role in PL comes from the fact that conditions where PL is high are also the regions where HP transfer is observed to be high (e.g., biodiversity hotspots; Alonso et al., 2010; Arceo‐Gómez et al., 2019) or expected to be high (urban or heavily invaded habitats; Johnson et al., 2019; J. M. Bennett et al., unpublished manuscript). If anthropogenic factors that reduce CP also lead to increased HP, e.g., loss of specialist pollinators and replacement by generalists (Johnson et al., 2019), then both processes may be acting in concert to reduce plant reproductive success, but without studying both aspects of pollen transfer we cannot know for certain.

Unfortunately, we find that despite the abundance of studies on plant reproductive ecology (Knight et al., 2018), there is a scarcity of cross talk between research on these two aspects of pollination. We recently reviewed, at a global scale, studies quantifying PL (1249 plant species from 927 studies; Bennett et al., 2018) and quantifying HP receipt (245 species from 26 studies; Arceo‐Gómez et al., 2019) and found when combining these data sets, measures of both entities are available for less than 1% of the plant species (Fig. 2). For these species, a few studies have measured both PL and HP receipt (Jakobsson et al., 2009; Montgomery, 2009; McKinney and Goodell, 2010), but only one study directly correlated HP receipt on PL (Waites and Agren, 2004). In comparing multiple populations of Lythrum salicaria , Waites and Agren (2004) found PL decreased with number of CP grains received but found no relation with number of HP grains.

The shortage of studies precludes a general test of the effect of HP receipt on PL or whether the HP effect depends on CP receipt or functional plant traits (e.g., anther or stigma placement) that modify HP receipt (i.e., reduce it), yet both warrant study because together they inform more fully on the biological mechanisms of PL. For instance, specialists (those pollinated by a single species or with restrictive morphologies) are the most pollen limited (J. M. Bennett et al., unpublished manuscript), but they are also expected to have the lowest HP (Arceo‐Gómez et al., 2016, 2019), so in these species PL is predicted to be related to low CP rather than high HP. In contrast, generalists (pollinated by multiple species, or with permissive morphologies) are less pollen limited (J. M. Bennett et al., unpublished manuscript), but have the greatest HP (Arceo‐Gómez et al., 2016, 2019), suggesting that they may have high CP and HP (unless CP delivery trades off with HP delivery). Thus, in these species, PL is predicted to be related to high HP rather than low CP. We acknowledge, however, that mechanisms to tolerate HP receipt (see Ashman and Arceo‐Gómez, 2013) may also exist and modify its effect on seed set.

In conclusion, determining whether conspecific and/or heterospecific pollen receipt lead to PL under anthropogenic change or contribute to variation in PL among plant species or phenotypes will require a pluralistic approach. To this end, we advocate for research that combines characterization of HP receipt when conducting supplemental pollinations to assess PL. It should include a characterization of both HP and CP on the stigmas of open‐pollinated and pollen‐supplemented plants to determine the potential contribution of HP receipt to PL. In addition, varying the timing of supplementation can be used to directly assess the impact of HP in the supplemented treatment (Fig. 1A, B). Furthermore, combining data on pollen transfer networks (i.e., amounts and identities of HP received) along with pollen supplementation studies of coflowering community members will provide insight into how pollinator‐mediated interspecific relationships contribute to PL variation within natural communities. Finally, hand pollinations can include different size loads of CP and HP and varying HP identity (see Arceo‐Gómez and Ashman, 2011) to identify specific effectors and their interactions. These data combinations will reveal the relative contribution of lower CP quantity and greater HP receipt to PL, and how each varies with plant traits and ecological factors. This mechanistic detail is increasingly important as different anthropogenic changes (species invasions/extinctions, land‐use conversion and climate change) may disrupt different underlying mechanisms of PL and do so differently across species. For instance, loss of specialist pollinators and rise of super generalists may lead to greater amounts of HP transfer (or from different HP donors than in the past; Johnson et al., 2019), and both HP donor identity and amount can affect reproductive success (Arceo‐Gómez and Ashman, 2011). Understanding these mechanisms will ultimately allow for more informed understanding of the drivers of PL as a result of anthropogenic change and lead to more effective plant conservation strategies. Further, because both HP and PL can determine species population growth and coexistence (Ashman et al., 2004; Ashman and Arceo‐Gómez, 2013; Schreiber et al., 2019), they may jointly be key to larger‐scale patterns of plant abundance and distribution.

人为的变化（物种入侵/灭绝，土地利用转换和气候变化）将如何影响世界被子植物的授粉和繁殖成功，其中85％需要动物授粉（Ollerton et al。，2011）？要回答这个问题，就需要了解引起种子生产花粉受限的机制。花粉限制（PL）广泛存在（Bennett等人，2018），并且发生在授粉媒介未能将足够数量或质量的花粉传授给柱头时（Ashman等人，2004））。据认为主要发生在传粉者很少去植物或传粉者没有带来足够的同种花粉（CP）或带来低质量的CP（例如，自花粉来自不同的花）上。相同的植物）（Ashman等人，2004； Aizen和Harder，2007）。然而，关于PL的讨论经常被遗忘的是，传粉媒介还可以在物种之间转移异种花粉（HP）（但请参见Jakobsson等，2009； McKinney和Goodell，2010）。

关于在花期后期采集的柱头上的HP收据的研究表明，几乎所有植物（88％）都至少获得了一定的HP，当它们这样做时，可以占总花粉量的75％（Ashman和Arceo-Gómez，2013 ; Arceo‐Gómez等，2019）。沉积在柱头上的HP可以干扰CP管的生长并减少CP的发芽和受精（Morales和Traveset，2008年），尽管其影响范围可能很大（Ashman和Arceo-Gómez，2013年））。因此，HP可以被视为花朵可能接受的最低质量的花粉，因为它不能导致合法的种子生产，甚至可以减少结实。此外，如果在传粉媒介的身上，HP和CP的运输之间需要权衡，那么更高的HP转移会导致较少的CP沉积在特定的柱头上（例如，Moreira-Hernández和Muchhala，2018年）。人为的环境变化可能会通过CP数量和质量的变化以及授粉媒介传递的HP的数量（和特性）的变化而影响PL。预测将来哪种植物物种最容易受到PL的侵害，需要对这两种机制都有所了解。

传统的PL测度方法包括随机分配花朵进行开放，自然授粉（对照）和手工授粉（补充）处理，然后测量补充处理中繁殖成功的程度（Bennett等，2018）。因为花粉补充处理适用于对自然授粉也开放的柱头（Ashman等人，2004 ; Bennett等人，2018），这些污名可能先前已经收到，或者随后可能同时收到了CP和HP。HP可以通过多种方式影响PL的估算。首先，补充处理中较高的繁殖成功表明PL，这可能不是由于CP不足，而是由于开放花中强烈的HP干扰（未补充）；CP补充剂治疗可增加CP与HP的比率，从而降低HP干扰的重要性（图1A）。另一种可能性是，补充处理中较高的繁殖成功可能是由于补充CP（在最近开放的花朵中）阻止了随后的HP沉积/附着，这在未补充CP的开放花朵中很常见（图1B）。 。持续，2018 ; 图1C）。在这些实验的极端情况下，与补充治疗相比，开放式治疗可能同时具有HP和低质量CP，补充治疗仅接受治疗中给出的高质量CP（套袋可防止其他任何东西到达柱头） 。值得注意的是，即使这些情况也过度简化了自然花粉转移的动力学。尽管如此，他们仍然揭示了授粉媒介体上花粉运输导致柱头（CP和HP之间）种间花粉相互作用的潜在复杂性（Arceo-Gómez和Ashman，2011 ; Minnaar等人，2018），并且可能影响PL 。

一个普遍的想法是植物进化为减少PL，PL是其进化稳定的最佳分配策略当代破坏的证据，这可能是由于传粉媒介的丧失和高质量CP的传递减少所致（Ashman等人，2004年综述））。的确，人为改变会导致严重的PL，特别是对于依赖传粉媒介的物种（Burns等人，2019; JM Bennett等人，未出版的手稿）。但是，我们不知道由于人为改变而导致的PL证据在多大程度上反映了CP数量减少，CP质量降低和/或HP接收量增加。HP接收量增加可能在PL中起重要作用的迹象是，PL较高的条件也是HP转移较高的区域（例如，生物多样性热点; Alonso等人，2010 ; Arceo- Gómez等人，2019年）或预计会很高（城市或严重入侵的栖息地； Johnson等人，2019年； JM Bennett等人，未发表的手稿）。如果降低CP的人为因素也导致HP升高，例如专业传粉媒介的丧失和通才的替代（Johnson等，2019），那么这两个过程可能会共同发挥作用，以减少植物繁殖的成功，但是如果不研究花粉转移的两个方面，我们就无法确定。

不幸的是，我们发现尽管对植物生殖生态学进行了大量研究（Knight等人，2018），但在授粉这两个方面的研究之间却缺乏相声对话。我们最近在全球范围内对量化PL（927个研究中的1249种植物; Bennett等人，2018）和量化HP接收量（来自26个研究中的245种;Arceo-Gómez等人，2019）进行了研究，并发现了何时结合这些数据集，两种实体的度量值不到1％的植物物种可用（图2）。对于这些物种，一些研究同时测量了PL和HP的接收量（Jakobsson等，2009； Montgomery，2009； McKinney和Goodell，2010）。），但只有一项研究直接关联了HP在PL上的收入（Waites和Agren，2004年）。在比较唾液千屈菜的多个种群时，Waites和Agren（2004）发现PL随接收到的CP粒数的减少而降低，但与HP粒数的数量没有关系。

缺乏研究无法对HP收据对PL的影响或HP效果是否取决于CP收据或能改变HP收据（即减少它）的功能性植物性状（例如花药或柱头放置）进行常规测试，但是两者值得研究，因为它们在一起可以更全面地了解PL的生物学机制。例如，专家（由单个物种授粉或具有限制性形态的传粉者）的花粉量最多（JM Bennett等人，未发表的手稿），但他们的HP也最低（Arceo-Gómez等人，2016年，2019年），因此在这些物种中，PL预计与低CP而不是高HP有关。与此相反，通才（由多个物种，或具有许可形态授粉）较少花粉限制（JM Bennett等人，未发表的手稿），但具有最大的HP（Arceo-Gomez等，2016，2019），这表明它们可能具有较高的CP和HP（除非CP交付需要与HP交付进行权衡）。因此，在这些物种中，预计PL与高HP而不是CP低有关。但是，我们承认，也可以存在耐受HP收据的机制（请参阅Ashman和Arceo-Gómez，2013年），并修改其对种子集的影响。

总之，确定同种和/或异种花粉的接收是否导致人为改变下的PL或导致植物物种或表型之间PL的变化，将需要采取多种方法。为此，我们提倡进行研究，以结合进行补充授粉评估PL时HP收据的特征。它应包括在开放授粉和补充花粉的植物的柱头上对HP和CP的表征，以确定HP收据对PL的潜在贡献。此外，可以通过改变补充时间来直接评估HP在补充治疗中的影响（图1A，B）。此外，结合花粉转移网络上的数据（即 共花社区成员的花粉补充研究，以及通过传粉媒介介导的种间关系如何促进自然社区内PL的变化，将有助于了解真相。最后，人工授粉可能包括不同大小的CP和HP负载以及不同的HP身份（请参见Arceo-Gómez和Ashman，2011年），以确定特定的效应子及其相互作用。这些数据组合将揭示较低的CP数量和较高的HP收据对PL的相对贡献，以及每种变化如何随植物性状和生态因素而变化。由于不同的人为变化（物种入侵/灭绝，土地利用转换和气候变化）可能破坏PL的不同潜在机制，并且跨物种的行为方式不同，因此，这种机械细节越来越重要。例如，专业传粉者的丧失和超级通才的崛起可能导致HP转移量的增加（或来自过去不同HP捐赠者的转移; Johnson等人，2019年），HP捐赠者的身份和数量都可能影响生殖成功（Arceo‐Gómez和Ashman，2011年）。理解这些机制最终将使人为变化导致对PL驱动因素的更充分的了解，并导致更有效的植物保护策略。此外，由于HP和PL均可确定物种种群的增长和共存（Ashman等人，2004； Ashman和Arceo-Gómez，2013； Schreiber等人，2019），它们可能共同成为更大规模植物格局的关键丰富和分布。