1

LRH: Márquez-Luna, Vázquez, Castellanos and Ortiz-Pulido 1

RRH: Hummingbird flower mite abundance 2

3

NUMBER OF HUMMINGBIRD VISITS DETERMINES FLOWER MITE ABUNDANCE ON 4

HUMMINGBIRD FEEDERS 5

Ubaldo Márquez-Luna1, 2

, María Magdalena Vázquez González3, Ignacio 6

Castellanos1, and Raúl Ortiz-Pulido

1* 7

8

1Centro de Investigaciones Biológicas, Instituto de Ciencias Básicas e Ingeniería, 9

Universidad Autónoma del Estado de Hidalgo, Km 4.5 Carretera Pachuca-Tulancingo, 10

Mineral de la Reforma, Pachuca, Hgo., C.P. 42001, México. 11

2Current address: Departamento de Biología, Universidad Autónoma Metropolitana, Av. 12

San Rafael Atlixco No. 186, Col. Vicentina, Delegación Iztapalapa, 09340 México, D.F. 13

3Departamento de Ciencias, División de Ciencias e Ingeniería, Universidad de Quintana 14

Roo, Boulevard Bahía s/n, esquina Ignacio Comonfort, Colonia del Bosque, Chetumal, 15

Quintana Roo, C.P. 77019, México. 16

*Corresponding author: raulortizpulido@yahoo.com, Phone +52 7717172 000 ext 6676, 17

Fax +52 7717172112 18

19

Revision accepted on April 2016 in Applied and Experimental Acarology 20

2

ABSTRACT 1

2

Members of several genera of mites from the family Melicharidae (Mesostigmata) use 3

hummingbirds as transport host to move from flower to flower, where they feed on pollen 4

and nectar. The factors that influence hummingbird flower mite abundance on host plant 5

flowers are not currently known. Here we tested whether hummingbird flower mite 6

abundance on an artificial nectar source is determined by number of hummingbird visits, 7

nectar energy content or species richness of visiting hummingbirds. We conducted 8

experiments employing hummingbird feeders with sucrose solutions of low, medium, and 9

high energy concentrations, placed in a xeric shrubland. In the first experiment, we 10

recorded the number of visiting hummingbirds and the number of visiting hummingbird 11

species, as well as the abundance of hummingbird flower mites on each feeder. Feeders 12

with the highest sucrose concentration had the most hummingbird visits and highest flower 13

mite abundances, however there was no significant effect of hummingbird species richness 14

on mite abundance. In the second experiment, we recorded flower mite abundance on 15

feeders after we standardized the number of hummingbird visits to them. Abundance of 16

flower mites did not differ significantly between feeders when we controlled for 17

hummingbird visits. Our results suggest that nectar energy concentration determines 18

hummingbird visits, which, in turn determines flower mite abundance in our feeders. Our 19

results do not support the hypothesis that mites descend from hummingbird nostrils more 20

on richer nectar sources; however, it does not preclude the possibility that flower mites 21

select for nectar concentration at other spatial and temporal scales. 22

Key words: host selection; nectar; Melicharidae; phoresis; hummingbirds. 23

24

3

HUMMINGBIRD FLOWER MITES (Acari: Mesostigmata: Melicharidae) comprise a well-known 1

ecological group that have many genera that lives, breeds, and feeds mostly on 2

ornithophilous host plants. They are found from USA to Argentina and Chile (Colwell and 3

Naeem 1994). Most species in this group are nectarivorous but also feed on pollen (Colwell 4

1973; Colwell 1985; Krantz and Lindquist 1979). 5

Most flower mites are associated with one or a few closely related plant species 6

(Colwell 1995; Heyneman et al. 1991; Naskrecki and Colwell 1998); however, there are 7

polyphagous mite species that use host plants from different orders or families (Colwell and 8

Naeem 1979; Colwell and Naeem 1994; García-Franco et al. 2001). During their life cycle 9

(7–12 days), mites can move around on a plant by walking to newly opened flowers 10

(Colwell 1985), or disperse on butterflies, bumblebees, bats or hummingbird pollinators to 11

colonize new hosts (Boggs and Gilbert 1987; Colwell 1985; Guerra et al. 2012; Proctor and 12

Owens 2000; Tschapka and Cunningham 2004). 13

In contrast to the host-plant specificity characteristic of most flower mite species 14

dispersed by hummingbirds, they may use any hummingbird species that visits their host 15

flowers (Heyneman et al. 1991). Mites climb to the beaks of these birds, and travel on the 16

surface or inside the nostrils where they remain until they descend to a new host plant 17

(Colwell 1973, 1995, Proctor & Owens 2000). Given that hummingbirds visit flowers for 18

only a few seconds, it has been hypothesized that flower mites must use precise cues that 19

stimulate them to leave a hummingbird individual and descend onto a suitable host flower 20

(Colwell 1995; Cutraro et al. 1998). 21

To date, the factors that determine hummingbird flower mite abundance on the 22

flowers of a host plant are not known. However, several factors have been shown to be 23

associated with mite flower choice; among these are flower microclimatic conditions 24

4

(Dobkin 1985), flower phenology (Dobkin 1984), and presence of nectar, pollen and sugar 1

(Colwell 1973; Cutraro et al. 1998; Heyneman et al. 1991), which has led to the suggestion 2

that olfactory signals may be used by mites to decide when to leave a hummingbird’s beak 3

(Colwell 1995; Cutraro et al. 1998; Heyneman et al. 1991). Given that several species of 4

nectarivorous arthropods are known to be able to distinguish the olfactory cues produced by 5

different nectar energy concentrations (Ayasse and Jarau 2014; Raguso 2008), it is 6

plausible that hummingbird flower mites may also be able to use olfactory information on 7

nectar energy content to decide how to distribute themselves among flowers; however, this 8

hypothesis has not been tested. 9

An understanding of the factors that determine the abundance of hummingbird 10

flower mites on flowers may contribute to our knowledge of the interactions among 11

pollinators, flower mites and plants. This is particularly important since these kind of 12

systems rely mainly on nectar as an energy source (Colwell 1995; Lara and Ornelas 2001) 13

and the depletion of such source can affect the biological interactions in the system. 14

However, it is not yet clear which what is the role of nectar energy concentration in the 15

functioning of such tripartite systems. 16

Hummingbirds evaluate a series of factors when selecting food, and nectar energy 17

concentration is one of these. It has been shown that hummingbirds forage at sites with 18

flowers that contain relatively high concentrations of sugar in their nectar (e.g., Hainsworth 19

and Wolf 1976). If hummingbirds preferentially visit plants with higher nectar sugar 20

content and mites use hummingbirds to move between plants, it is likely that plants with 21

higher nectar sugar concentrations will be visited more frequently by hummingbirds, and 22

thus will have higher mite abundances than plants with lower sugar concentrations. 23

Moreover, we hypothesize that an increase in the number of hummingbird species visiting 24

5

nectar sources will increase the possibility that more flower mites reach the nectar source. It 1

is because: 1) not all hummingbird species visit all plant species at a particular site, 2) mites 2

use just a small subset of plant species available at a particular site and 3) mites use any 3

hummingbird species that visit the flowers where they are. Thus, an increase in the number 4

hummingbird species visiting a feeder increases the probability that at least one of them 5

encountered mites and hence can deposit them on the feeder. Thus, we hypothesize that the 6

number of flower mite individuals in an artificial nectar source will be determined by the 7

number of hummingbird visits, the energy content of the nectar and the number of visiting 8

hummingbird species. Thus, in this study, we set up two alternative hypotheses: abundance 9

of flower mites should be higher on high concentration nectar sources because 1) flower 10

mites prefer this kind of source and have a higher probability of descending from the 11

hummingbird if the nectar has a high concentration of sugar (i.e. direct selection of the 12

source), or; 2) these sources receive more hummingbird visits and the mites show no 13

differential tendency to descend based on sugar concentration (i.e. indirect selection of the 14

source mediated by hummingbirds). To test these hypotheses, we used two experiments in 15

which we placed artificial hummingbird feeders with sugar solutions of different energy 16

concentrations in the field. The employment of artificial feeders facilitates counting mites 17

and manipulating nectar concentration, and, since pollen is not included, mite decisions will 18

be influenced only by the factors considered in our hypothesis. 19

20

METHODS 21

Field work was carried out in a xeric shrubland site (98°42’42.12’’W - 20°08’03.71’’N; 22

2400-2650 m asl) close to the city of Pachuca, Hidalgo, Mexico. Field work was conducted 23

in October 2012 (experiment 1) and October 2014 – February 2015 (experiment 2), periods 24

6

when ornithophilous plants reduce their flower production in the site (Díaz-Valenzuela 1

2008). The plants that bloom during this period and visited by hummingbirds are: 2

Bouvardia ternifolia (Rubiaceae), Castilleja tenuiflora, Penstemon barbatus, Lamouroxia 3

dasyantha (Scrophulariaceae), Salvia prunelloides, Salvia amarissima (Lamiaceae), 4

Cuphea aequipetala (Lythraceae), Silene laciniata (Caryophyllaceae), Opuntia sp. 5

(Cactaceae) and Eysenhardttia polystachya (Leguminosae) (Díaz-Valenzuela 2008). During 6

this time, several hummingbird species, including Calothorax lucifer, Hylocharis leucotis, 7

Selasphorus platycercus, S. rufus, Archilochus colubris and Colibri thalassinus, are present 8

at our study site (Díaz-Valenzuela 2008; Mauricio-López 2005). 9

10

EXPERIMENTAL PROTOCOL.- We placed nine hummingbird feeders on a 300 300 m grid, 11

placing each feeder on top of a 1 m plastic pole, 100 m away from the other feeders. During 12

the field work, all feeders contained at least 100 ml of artificial nectar (sugar water) with 13

one of three sucrose concentrations (10, 20, and 30%) representing low medium and high 14

energy concentrations. These concentrations were intended to represent the range of nectar 15

sugar concentrations available in most of the flowers pollinated by hummingbirds (i.e., 7.3–16

66.5% total sugar, but the concentrations below 35% are the most common) (Pyke and 17

Waser 1981). We assigned the three sucrose concentrations randomly to the different 18

feeders. 19

Before each observation period, we cleaned the feeders to remove any mites present. 20

In order to prevent any walking mites from reaching the feeders, we applied 5-10 cm of 21

Tanglefoot (sticky resin; The Tanglefoot Co., Grand Rapids, MI, USA) at the poles that 22

7

supported the feeders. The application was always above the vegetation level, commonly 1

50-70 cm above the ground. 2

We observed each feeder with binoculars (842 mm) from 10 m away for 30-60 3

minutes (see below). We randomized the order in which we observed feeders. Field 4

observations were done from 0700 to 1300 h. We used field guides to identify the 5

hummingbird species (Howell 2003; Ortiz-Pulido et al. 2008; Van Perlo 2006). At the end 6

of each observation period, all mites were removed from the feeder and placed in 70% 7

ethanol. We identified the composition of the mite samples to the genus level using 8

specialized keys (Naskrecki and Colwell 1998) and by consulting a taxonomic specialist 9

(Gilberto José de Moraes, Universidade de Sao Paulo). 10

We ran two experiments. In experiment 1, we filled each feeder with its assigned 11

sugar concentration, waited at least one day, then cleaned the feeder and changed nectar 12

and observed each feeder for a period of 30 minutes per day every day for 7 days in 13

October 2012. We recorded the identity of visiting hummingbirds, number of hummingbird 14

visits per species, and number of flower mites on the feeder at the end of the 30 minute 15

period. 16

In experiment 2, we controlled the number of hummingbird visits by terminating 17

our observations after five hummingbird visits to each feeder. We did not determine if the 18

visits were done by the same or different hummingbird individual. However, sometimes we 19

saw two or three hummingbird individuals fighting for access to the feeder. If we did not 20

observed five visits within a 60 minute period we cleaned the feeder and restarted our 21

observations. This experiment was conducted on the same 300 x 300 grid as in the previous 22

experiment and also at two nearby (< 3 km) sites. In this experiment we filled each feeder 23

8

with a 20% sucrose concentration solution, waited at least one day, and then cleaned the 1

feeder and changed the concentration just prior to beginning our observations. We recorded 2

the same information as in experiment 1, except hummingbird richness because it did not 3

had significant effect in the first experiment (see results section). 4

5

STATISTICAL ANALYSIS.- For experiment 1, we analyzed mite abundance as a function of 6

the number of hummingbird visits, nectar sugar concentration, number of visiting 7

hummingbird species, date of sampling and feeder using a generalized linear model (GLM). 8

We included date and feeder to account for pseudoreplication, i.e. observing each feeder 9

seven times (each time in a different hour of the day but in different day). We included 10

interaction terms for all variables, except date and feeder. For experiment 2, we used a 11

GLM to analyze mite abundance as a function of nectar sugar concentration. 12

At all nectar concentrations, the response variable (mite abundance) had an 13

exponential distribution, which was verified with a χ2 goodness of fit test (in all the cases 14

χ2<7.1, df=5, P>0.05). GLM´s were done using GenStat (VSN_International 2011), 15

implicitly declaring the exponential distribution. The resulting variance revealed that the 16

data did not exhibit overdispersion; thus a posteriori adjustments were not performed 17

(Crawley 1993).When significant differences were found, a Tukey test for a posteriori 18

comparisons was used to determine which treatments differed significantly (Zar 1984). 19

20

RESULTS 21

Only adult mite individuals of one morphospecies of the genus Tropicoseius (Melicharidae) 22

were found on the feeders on both experiments. Male-female sexual proportion varied 23

between 1:4 to 1:2.5 (deviate to females) and 20-37% of females carried eggs. 24

9

For experiment 1 a total of 140 adult mite individuals arrived to feeders during a 1

31:30 hrs period. The number of arrived mites in 30 minutes was positively related to the 2

number of hummingbird visits (χ2 = 18.29, df = 1, 39, P < 0.001; Fig. 1). There were no 3

other significant relationships (Table 1). Feeders with the highest sugar concentration 4

(30%) were visited by hummingbirds significantly more (102 visits) than feeders with 5

medium (61 visits) and low (45 visits) concentrations (10% vs 30%: Q = 5.085, P = 0.001; 6

20% vs 30%: Q = 3.658, P = 0.032; Table 2, Fig. 2a). We recorded a total of five 7

hummingbird species visiting the feeders (Table 2). Feeders with 30 and 20% nectar sugar 8

concentrations were visited by five hummingbird species, while those with 10% were 9

visited by four species (Table 2). Most visits to feeders were by Calothorax lucifer (138 10

visits), while the species with the lowest number of visits was Cynanthus latirostris (two 11

visits; Table 2). 12

Even though sugar concentration was not significant in the overall GLM (Table 1), 13

more mites arrived on feeders with 30% sucrose than on feeders with 10% sucrose (Q = 14

3.489, P = 0.01; Fig. 2b). 15

In the second experiment, where we controlled for the number of hummingbird 16

visits, we found that 33 flower mites arrived on the feeders. On average 0.15 ±0.17 17

(mean±1 s.e.) flower mites were left per hummingbird visit on each feeder (n=215 18

hummingbird visits). The number of flower mites arriving did not significantly differ 19

among nectar concentrations (χ2 = 1.10, df=1,42, P=0.30). 20

21

DISCUSSION 22

The results obtained in our study site support the hypothesis that abundance of the 23

Tropicoseius species studied is higher on high concentration nectar sources because these 24

10

sources receive more hummingbird visits than low concentration nectar sources. To our 1

knowledge it is the first time that such a relationship has been reported to a flower mite 2

species. 3

The fact that we show that nectar concentration only indirectly affects mite 4

abundance contradicts other studies that suggest that nectar might directly affect mite 5

abundance. For example, it has been shown that hummingbird flower mites prefer sucrose 6

water than plain water (Colwell 1973). In addition, it has been shown that mites are able to 7

discriminate between nectar from their own host flower species and nectar from other 8

flowers species, and that they prefer nectar from their host flowers species (Cutraro et al. 9

1998; Heyneman et al. 1991). The conflicting results among studies could be explained by 10

the difference in scales over which mite flower selection was assessed. In the previous 11

studies, flower mites were unaided by hummingbird flight and thus restricted to moving at 12

most just several millimeters by, at least, 1 hour, i.e. the mite has enough time to choose its 13

way. On the other hand, in our study flower mites were carried by hummingbirds over tens 14

or hundreds of meters and they had but just a few seconds to choose if they must leave the 15

hummingbird bill. Thus, our results do not necessarily reject the hypothesis that flower 16

mites respond to flower signals to make decisions. It is because, in addition to nectar 17

concentration, there are other features of real flowers, such as fragrance and surface 18

chemicals (Heyneman et al. 1991), that could allow mites to make choices. As it has been 19

suggested (Colwell 1995), flower mites traveling in the hummingbird nostrils are exposed 20

to an air flux that contains chemical information of the environment near to the 21

hummingbird bill, and the flower mites could use it to decide to disembark form the 22

hummingbird bill. Rather, our results only show that the nectar concentration does not 23

11

explain by itself abundance of disembarked mites when mites are using hummingbirds to 1

move between/to feeders. 2

Our results show that dispersal or colonization by mites in our feeders is directly 3

affected by hummingbird visits. The relationship between the number of hummingbird 4

visits and mite abundance on feeders can be explained if we consider: 1) the number of 5

mites that can be carried by a hummingbird individual and 2) that an increase in 6

hummingbird visits will result in more mites reaching a host plant flower. It is known that 7

hummingbirds can carry up to 32 mites in their nostrils (Colwell 1973) and flowers may 8

contain up to 152 mites (Colwell 1995). Therefore, it is reasonable to consider that if a 9

nectar resource receives more hummingbird visits, the probability of more mites reaching 10

this resource increases, and therefore mite abundance is likely to be higher than at a less 11

visited nectar resource. However, such relationship has not been tested in nature and it 12

could change, because number of mites on flowers could be related to intrinsic features of 13

mite population dynamics within flowers (e.g. birth and death). Besides, flower mites can 14

both immigrate and emigrate from flowers using hummingbirds, and therefore, number of 15

hummingbird visits can also decrease number of flower mite in flowers. 16

In conclusion, we found that feeders with the highest nectar sugar concentration 17

received more hummingbird visits, and that, in turn, more hummingbird visits resulted in a 18

higher abundance of flower mites. In a natural system, our findings could imply that a plant 19

species that produces more nectar to attract more visitors to increase its pollination success 20

could in turn increase the number of mites on its flowers; to date, it is unclear if this could 21

diminish the plant’s reproductive success (Lara and Ornelas 2001). These implications 22

require explicit testing. 23

24

12

ACKNOWLEDGMENTS 1

2

We thank the Universidad Autónoma del Estado de Hidalgo for its support, Carlos Lara for 3

his comments during the beginning of this project, Nico Bluthgen for suggest the second 4

experiment, M. Isabel Herrera Juárez and Maria de los Angeles Soto Pineda for field 5

assistance, Richard Feldman for comments on an early version of this paper, Gilberto José 6

de Moraes, Departmento of Entomología e Acarologia, Escola Superior de Agricultura 7

"Luiz de Queiroz", University of Sao Paulo, Piracicaba, Sao Paulo, Brazil for mite 8

identification, and Richard Feldman and Margaret Schroeder for their help improving the 9

English. UML thanks Consejo Nacional de Ciencia y Tecnología for grant no. 266931. We 10

are also grateful to Fondos Mixtos-Consejo Nacional de Ciencia y Tecnología Hidalgo 11

project 191908 “Diversidad Biológica del Estado de Hidalgo (tercera etapa)”, and Consejo 12

Nacional de Ciencia y Tecnología project 161702 “Mejoramiento y actualización de la 13

infraestructura experimental para proporcionar soporte a los posgrados en Biodiversidad y 14

Conservación de la Universidad Autónoma del Estado de Hidalgo” for their generous 15

support. 16

Conflict of interest: the authors declare that they have not conflict of interest. 17

Ethical approval: All applicable international, national, and/or institutional guidelines for 18

the care and use of animals were followed. This article does not contain any studies with 19

human participants performed by any of the authors. 20

21

13

Figure legends 1

2

Fig. 1. Relationship between the number of hummingbird visits and the number of flower 3

mites on feeders placed in a xeric shrubland outside of Pachuca, Hidalgo, Mexico. Solid 4

line displays the regression line and dashed lines show the 95% confidence limits of its 5

predicted distribution 6

7

Fig. 2. Mean number (± s.e.) of (A) hummingbird visits and (B) hummingbird flower mites 8

on feeders for each of the three nectar sugar concentrations used in this study. Different 9

lower case letters indicate significant differences (P < 0.05) 10

11

12

14

Figure 1. (Program computer used to create the figure: SigmaStat) 1 2

Hummingbird visits

0 2 4 6 8 10 12

Nu

mb

er

of

flo

we

r m

ite

s

-2

0

2

4

6

8

10

3 4

15

Figure 2. (Program computer used to create the figure: SigmaStat) 1

2

3

Vis

itin

g h

um

min

gb

ird

s

0

1

2

3

4

5

6

Sugar concentration

Nu

mb

er

of

mite

s

0

1

2

3

4

10 20 30

a

a,b

b

a

a

b

A

B

16

1

Table 1. Generalized linear model testing the significance of number of visiting

hummingbirds (number of visits), nectar energy content (sugar concentration), visiting

hummingbird species richness, date of sampling and feeder, on hummingbird flower mite

abundance.

Factor df Deviance χ2 P

Number of visits 1 18.29 18.29 <0.001

Sugar concentration 2 0.09 0.05 0.954

Species richness 1 0.08 0.08 0.774

Number of visits Sugar concentration 2 0.04 0.02 0.980

Number of visits Species richness 1 0.29 0.29 0.592

Sugar concentration Species richness 2 1.19 0.60 0.552

Sugar concentration Number of visits Species

richness

2 0.19 0.10 0.909

Date of sampling 6 1.09 0.18 0.982

Feeder 6 2.54 0.42 0.864

Residual 39 8.85

2

3

4

17

Table 2. Number of visits per hummingbird species

categorized by nectar sugar concentration. Data from

experiment 1.

Species

[ ] sugar

10% 20% 30%

Calothorax lucifer 27 49 62

Selasphorus platycercus 10 6 15

Selasphorus rufus 7 2 13

Eugenes fulgens 1 3 11

Cynanthus latirostris 0 1 1

1

2

18

LITERATURE CITED 1

Ayasse, M., S. Jarau (2014) Chemical ecology of bumble bees. Annu Rev Entomol 59, 299-319 2 Boggs, C.L., L.E. Gilbert (1987) Spatial and temporal distribution of lantana mites phoretic on 3

butterflies. Biotropica 19, 301-305 4 Colwell, R.K. (1973) Competition and coexistence in a simple tropical community. American 5

Naturalist 107, 737-760 6 Colwell, R.K. (1985) Stowaways on the hummingbird express. Nat Hist 94, 56–63 7 Colwell, R.K. (1995) Effects of nectar consumption by the hummingbird flower mite proctolaelaps 8

kirmsei on nectar availability in hamelia patens. Biotropica 27, 206-217 9 Colwell, R.K., S. Naeem (1979) The first known species of hummingbird flower mite north of 10

mexico: Rhinoseius epoecus n. Sp. (mesostigmata: Ascidae). Ann Entomol Soc Am 72, 485-11 491 12

Colwell, R.K., S. Naeem (1994) Life history patterns of hummingbird flower mites in relation to host 13 phenology and morphology. In: Houck, M.A. (ed) Mites: Ecological and evolutionary 14 analyses of life history patterns.(pp23-44). New York: Chapman and Hall 15

Crawley, M.J. (1993) Glim for ecologist. London, Blackwell Scientific Publications 16 Cutraro, J.L., A.Y. Ercelawn, E.G. LeBrun, E.W. Lonsdorf, H.A. Norton, M.J. McKone (1998) 17

Importance of pollen and nectar in flower choice by hummingbird flower mites, 18 proctolaelaps kirmsei (mesostigmata: Ascidae). Int J Acarol 24, 345-351 19

Díaz-Valenzuela, R. (2008) Análisis descriptivo del sistema colibrí-planta en tres niveles de las 20 escalas espacial, temporal y en la jerarquía ecológica en un paisaje mexicano. Master, 21 Centro Iberoamericano de la Biodiversidad, Instituto de Ecología, Universidad Autónoma 22 del Estado de Hidalgo 23

Dobkin, D.S. (1984) Flowering patterns of long-lived heliconia inflorescences: Implications for 24 visiting and residente nectarivores. Oecologia 64, 245-254 25

Dobkin, D.S. (1985) Heterogeneity of tropical floral microclimates and the response of 26 hummingbird flower mites. Ecology 66, 536-543 27

García-Franco, G.J., D.M. Burgoa, T.M. Pérez (2001) Hummingbird flower mites and tillandsia spp. 28 (bromeliaceae): Polyphagy in a cloud forest of veracruz, mexico. Biotropica 33, 538-542 29

Guerra, T.J., G.Q. Romero, J.C. Costa, A.C. Lofego, W.W. Benson (2012) Phoretic dispersal on 30 bumblebees by bromeliad flower mites (mesostigmata, melicharidae). Insect Soc 59, 11-16 31

Hainsworth, R.F., L.L. Wolf (1976) Nectar characteristics and food selection by hummingbirds. 32 Oecologia 25, 101-113 33

Heyneman, A.J., R.K. Colwell, S. Naeem, D.S. Dobkin, B. Hallet (1991) Host plant discrimination: 34 Experiments with hummingbird flower mites. In: Price, P.W., T.M. Lewinsohn, G.W. 35 Fernandes, W.W. Benson (eds) Plant–animal interactions: Evolutionary ecology in tropical 36 and temperate regions.(pp455-485). New York: John Wiley and Sons 37

Howell, S.N.G. (2003) Hummingbirds of north america the photographic guide. Princeton, NJ., 38 Princeton University Press 39

Krantz, G.W., E.E. Lindquist (1979) Evolution of phytophagous mites (acari). Annu Rev Entomol 24, 40 121-158 41

Lara, C., J.F. Ornelas (2001) Nectar ‘theft’ by hummingbird flower mites and its consequences for 42 seed set in moussonia deppeana. Funct Ecol 15, 78–84 43

Mauricio-López, E. (2005) Interacción colibrí-planta: Variación espacial en un matorral xerófito de 44 hidalgo, méxico. Universidad Autónoma del Estado de Hidalgo 45

19

Naskrecki, P., R.K. Colwell (1998) Systematics and host plant affiliations of hummingbird flower 1 mites of the genera tropicoseius baker & yunker and rhinoseius baker & yunker (acari: 2 Mesostigmata: Ascidae), Entomological Society of America 3

Ortiz-Pulido, R., E. Mauricio-Lopéz, V. Martínez-García, J. Bravo-Cadena (eds)(2008) Sabes quien 4 vive en el parque nacional el chico? Colibríes. Universidad Autónoma del Estado de 5 Hidalgo, Pachuca, Hidalgo, México 6

Proctor, H., I. Owens (2000) Mites and birds: Diversity, parasitism and coevolution. Trends Ecol 7 Evol 15, 358-364 8

Pyke, H.G., N.M. Waser (1981) The production of dilute nectars by hummingbird and honeyeater 9 flowers. Biotropica 13, 260-270 10

Raguso, R.A. (2008) Wake up and smell the roses: The ecology and evolution of floral scent. Annu 11 Rev Ecol Evol Syst 39, 549-569 12

Tschapka, M., S.A. Cunningham (2004) Flower mites of calyptrogyne ghiesbreghtiana (arecaceae): 13 Evidence for dispersal using pollinating bats. Biotropica 36, 377-381 14

Van Perlo, B. (2006) Birds of mexico and central america. Princeton, New Jersey, Princeton 15 University Press 16

VSN_International (2011) Genstat for windows. vol 14th. Hemel Hempstead, UK: VSN International 17 Zar, J.H. (1984) Biostatistical analysis, Second edn. Englewood Cliffs, New Jersey, Prentice-Hall 18

19