1 Biology Department, Concordia University, Montreal, Canada.
2 Red de Biología Evolutiva, Instituto de Ecología A. C., Xalapa, México.
3 MARE Departamento de Ciências da Vida, Universidade de Coimbra, Coimbra, Portugal.
4 Departamento de Biología Celular y Fisiología, Unidad Foránea Tlaxcala, Instituto de Investigaciones Biomédicas, Universidad Nacional Autónoma de México, Tlaxcala, México.
5 The New Zealand Institute for Plant and Food Research Ltd, Auckland, New Zealand.
6 School of Biological Sciences, University of Auckland, Auckland, New Zealand.
7 Department of Biology, Lund University, Lund, Sweden.

Theoretical work suggests that reinforcement can cause the strengthening of prezygotic isolation in sympatry by mitigating the costs of maladaptive hybridization. However, only a handful of studies have simultaneously tested multiple predictions of this theory in natural populations. We investigated reinforcement in a mottled hybrid zone between the damselflies Ischnura elegans and I. graellsii, which are characterized by incomplete and asymmetric reproductive isolation and exhibit reproductive character displacement in mating-related structures. We tested the conditions for reinforcement by quantifying whether hybridization was costly and prezygotic isolation stronger in sympatry compared with allopatry. Additionally, we investigated two specific predictions of reinforcement: i) greater premating asymmetries in sympatry; and ii) weaker postzygotic isolation in sympatry than in allopatry. Our findings indicate the presence of maladaptive hybrids, which suggests Bateson-Dobzhansky-Müller incompatibilities in allopatry. We also found that reinforcement has strengthened mechanical isolation, at least in one direction in sympatry. We observed evidence for greater premating asymmetries in sympatry than in allopatry, which is consistent with reinforcement. However, fully testing the prediction of weaker postzygotic isolation in sympatry compared to allopatry was hindered by the highly asymmetrical levels of reproductive isolation between the two reciprocal cross directions. Our study highlights a case where reinforcement and heterospecific gene flow exert opposite effects on reproductive isolation between reciprocal crosses, where reinforcement increases reproductive isolation in one direction while gene flow weakens it in the opposite direction.