Ctenocephalides felis felis vs . Ctenocephalides canis ( Siphonaptera : Pulicidae ) : some issues in correctly identify these species Ctenocephalides felis felis vs . Ctenocephalides canis : ( Siphonaptera : Pulicidae ) : algumas questões para identificar corretamente estas espécies

Ctenocephalides felis felis is one of the most important ectoparasites of dogs and cats throughout the world, because of its geographical distribution, dual parasitological action as an infesting agent and a vector of diseases, the economic losses and the acquired resistance against common insecticides. In Brazil, it surpasses Ctenocephalides canis in distribution, number of host species infested, prevalence and epidemiological importance. However, in some studies the species have been misidentified on the basis of their morphological characters included in taxonomic keys. The morphological variations of chaetotaxy, especially those on the dorsal margin of the hind tibia and lateral metanotal area (LMA), found in certain specimens, have sometimes been erroneously treated as hybrids, in spite of the nonexistence of the two species of Ctenocephalides in the same municipality or region. This review focuses on the characteristics used for interspecific diagnosis and intraspecific variations found between the species. Data on distribution, hosts, prevalence and parasitological action are also presented as an auxiliary means for recognizing the species.


Introduction
Within the family Pulicidae of the order Siphonaptera, the genus Ctenocephalides Stiles & Collins, 1930, includes 13 species and subspecies (BEAUCOURNU; MÉNIER, 1998), but only two, respectively the dog flea Ctenocephalides canis (Curtis, 1826) and the cat flea Ctenocephalides felis felis (Bouché, 1835), are cosmopolitans that have been recorded in South America.C. felis felis is more adaptable than C. canis, since it infests more host species and has therefore become established in more extensive areas (HOPKINS et al., 1953).According to KOUTINAS et al. (1995), the distribution of these species is related to environmental factors that influence their survival, development and reproduction.
Although the two species of Ctenocephalides can occur in the same Brazilian geographical region and infest the same host species, C. felis felis is more important in disease transmission.Moreover, C. felis felis has also already shown resistance to common insecticides for two decades (EL-GAZZAR et al., 1986).For this reason, correct species identification is essential for control measures and genetic inferences.
The two species are usually separated according to the shape of the head, length of the first spine of the genal comb, number of bristles on the lateral metanotal area (LMA) (Figures 1, 2) and the number of short stout bristles in the interval between the postmedian and apical long bristles of the dorsal margin of the hind tibia (HOPKINS et al., 1953;JOHNSON, 1957;AMIN, 1976;MÉNIER;BEAUCOURNU, 1998;BEAUCOURNU;MÉNIER, 1998;LINARDI;GUIMARÃES, 2000) (Figures 3, 4; Table 1).Males can be further identified by the shape of the manubrium of the clasper (HOLLAND, 1949) (Figures 5, 6) and the size of the hamulus on the aedeagus (Figures 2 and 3 of MÉNIER; BEAUCOURNU, 1998).However, in spite of these differences, some variations in the chaetotaxy and the number of spines in the genal comb have been found (AMIN, 1976;AMIN et al., 1974).Some misidentifications have also been noticed: in the Afrotropical zone, C. felis strongylus has often been identified incorrectly as

C. felis felis C. canis Shape of the head
Length generally greater than twice the height (Figure 2A   C. canis based on the sole criterion of the cephalic profile; and likewise, C. orientis in Asia (BEAUCOURNU; KOCK, 1990).
Ctenocephalides collected from epidemiological investigations, particularly in relation to plague in the Palaearctic and Nearctic regions, were often named according to their hosts, although urban dogs are much more infested with C. felis felis than with C. canis (BEAUCOURNU, 1973).On the other hand, some characteristics have been improperly interpreted when using certain taxonomic keys.

Hosts of Ctenocephalides
Ctenocephalides species infest carnivores, edentates, lagomorphs, marsupials, primates, rodents and ungulates, and can also be found in the nests, burrows, trails and tracks of their hosts, as well as indoors, where they colonize profusely (LINARDI; GUIMARÃES, 2000).Brazilian host species recorded for species of Ctenocephalides include seven orders and 41 species of mammals, as well as one species of bird infested by C. felis felis (Table 2).Although carnivores can be considered to be the true or primary hosts, infestations on Brazilian carnivores and rodents represent respectively 26.8% and 43.9% of the findings.The cat flea is commonly collected on opossums because of the varied habitats that they use (LINARDI, 2006).On the contrary, C. canis is only seen on domestic carnivores.

Ctenocephalides Species as Vectors of Diseases
Amongst the arthropods, C. felis felis is the most important ectoparasite of dogs and cats throughout the world, causing annoyance to the animals and acting as a vector of diseases (RUST;DRYDEN 1997).It causes allergic dermatitis (HALLIWELL, 1979) and in heavy infestations can also cause iron deficiency anemia in young animals (HARVEY et al., 1982).More than twenty different    types of endosymbionts or pathogens have been found to be associated with species of Ctenocephalides as biological vectors or intermediate hosts, including bacteria, protozoa and helminths, thus representing a potential health risk for humans (JELLISON, 1959;JENKINS, 1964;BEARD et al., 1990;KRÄMER;MENCKE, 2001;LINARDI;GUIMARÃES, 2000;LINARDI, 2001).The cat flea also lodges other monoxenic organisms, such as gregarines, microsporidians and trypanosomatids.The endosymbionts or pathogens found in Ctenocephalides species in Brazil are shown in Table 3.According to De Avelar et al. (2007Avelar et al. ( , 2008)), these endosymbionts might be useful for biological control of C. felis felis in environments exhibiting high levels of infestation.Also using molecular techniques, Coutinho and Linardi (2007) and Ferreira et al. (2009a) recognized DNA of Leishmania chagasi in C. felis felis, thus opening new perspectives for mechanical transmission of canine visceral leishmaniasis.
De Avelar et al. ( 2011) raised the possibility that Leptomonas of fleas may be pathogenic to humans and dogs because some people crush dog fleas between their fingers and carry them to their mouths.Besides, dogs are likely to ingest flea gut contents or entire fleas when crushing the insects between their teeth or licking their own fur or that of other dogs.According to Garin et al. (2001), it is possible that the numbers of human cases of infection with lower trypanosomatids is underestimated because of the morphological similarity to and cross-reactivity with Leishmania infantum chagasi.
In addition to the role as a transmitter of diseases, C. felis felis provokes allergic dermatitis and has been reported to produce anemia in dogs, cats, goats, cattle and sheep (OBASAJU; OTESILE, 1980;YERUHAM et al., 1989).
In Brazil, C. felis felis is also the most common flea of Canidae and has been found in 17 states: Alagoas, Amazonas, Bahia, Ceará, Espírito Santo, Goiás, Mato Grosso, Minas Gerais, Paraíba, Paraná, Pernambuco (including the territory of Fernando de Noronha), Rio de Janeiro, Rio Grande do Norte, Rio Grande do Sul, Roraima, Santa Catarina and São Paulo.On the other

Morphological Variations
In Ctenocephalides spp., the most frequent morphological variations are observed in the combs and chaetotaxies of LMA (erroneously referred to as the metepisternum by some authors) and in hind tibia (AMIN et al., 1974;AMIN, 1976).In Canada, Holland (1949) found certain nontypical specimens, usually females, which were difficult to identify with certainty.Because of this, the two species were treated collectively in those localities where they were collected.
Alterations in chaetotaxy on the LMA and metatibia might suggest hybridization between the two species, as previously proposed by Holland (1949), Fox (1952), Amin et al. (1974) and Amin (1976).According to Benton (1998), hybrids depend upon occurrence of two closely related species in close association, such as in fleas of the genus Ceratophyllus Curtis, 1832, which share the same bird nests in North America.However, the hypothesis of hybridization between C. felis felis and C. canis must be rejected because good species do not cross with each other, as reinforced by Beaucournu and Guiller (2006).
In Brazil, altered chaetotaxies on the LMA or hind tibia of Ctenocephalides spp.were also observed by Nagem (1977) when examining fleas from dogs in the municipality of São João d'El Rei, state of Minas Gerais, and by Fernandes et al. (1996) in fleas from dogs and cats in the municipality of Rio de Janeiro, state of Rio de Janeiro.However, in these surveys, they were named C. felis felis.Other variations concerning the chaetotaxy of these species were unduly designated hybrids in some studies (RODRIGUES et al., 2008;SANTOS, 2008;STALLIVIERE et al., 2009).
In a sample of 87 fleas collected from 33 dogs in the rural zone of the municipality of Jaboticatubas, around Serra do Cipó National Park (19° 30' S and 43° 44' W), state of Minas Gerais, Brazil, Santos (2008) found that all the dogs were infested by C. f. felis, and 18 of them (54.5%) by specimens presenting variations in the number of bristles on the LMA and/or metatibia.Out of the 87 specimens of C. f. felis examined, 27 (31.0%)presented variations both on the LMA and hind tibia (Table 5).Separately by sex, the variations represented 40.7% in females and 15.1% in males, with a significant difference between them when compared using the chi-square test (χ 2 = 6.26; p < 0.05).Also in this sample, 15 specimens (17.2%) presented variations only on one side (asymmetrical), while in four (4.6%) the alterations were observed on both sides of the thorax (symmetrical) (Table 6).About 21.8% of the alterations from the standard numbers consisted of an increase of at least one bristle.Alterations occurred significantly more frequently in females (33.3%) than in males (6.1%) (χ 2 = 7.1; p < 0.01).Figure 7 shows the variation of the number of bristles on the left side of the LMA.The variations of chaetotaxy of the hind tibia among the sampled specimens of C. f. felis are indicated in Table 7; 12.7% of the specimens examined presented such variations.Figure 8 illustrates one of these variations.Contrary to the characteristics of the LMA, the chaetotaxy of the hind tibia was approximately equally variable in both sexes.Asymmetrical variations (9.2%) were more frequent than symmetrical ones (3.4%).Likewise, variations involving increased numbers of bristles were more frequently found (9.2%) than ones relating to decreased numbers (3.4%) (Table 7).
Although the total numbers of variations (Table 5) and the ones concerning the LMA (Table 6) in the present study were more significant in females, and a great number of females were collected (62.1%),Amin (1976) found similar results when studying dog and cat fleas from Wisconsin, USA.Interestingly, both variations of bristles on the LMA (Table 6) and hind tibia (Table 7) were more significant in relation to increased numbers of bristles, rather than decreased numbers.

Final Remarks
Given the enormous epidemiological and economical importance of these fleas, correct identification of species is indispensable.Pet flea-related diseases account for over 50% of the dermatological cases reported to veterinarians (BEVIER-TOURNAY, 1989).Moreover, the annual expenditure by pet owners on flea control products in the United States exceeds $ 1 billion (CONNIFF, 1995).Besides the economical aspect, C. felis felis has been reported to have developed resistance to at least five different categories of insecticides, including carbamates, organophosphates, pyrethroids, pyrethrins and organochlorines (WHO, 1992).
In Brazil, C. felis felis is the most important flea species on pets because of its geographical distribution, number of other hosts parasitized and vector competence.It is more euryxenous than C. canis, having been recorded on 41 different host species (Table 2) and 17 states, as well as being found naturally infected by eight different species of endosymbionts (Table 3).The infestations are more prevalent in southeastern states (Table 4).In all states, cats exclusively harbored C. felis felis.Except for Manaus and Salvador, where only small numbers of few dogs were sampled, the prevalence of infestation by C. canis seems to increase from north to south.Interestingly, Castro and Rafael (2006) found no specimens of C. canis on dogs in Manaus, whereas previously Gordon and Young (1922) observed dogs infested only by this species.It remains unclear as to whether this finding was geographically isolated or whether, in fact, this species is truly decreasing in numbers towards extinction; or else, whether it was incorrectly identified.
It is important to stress that the head curvature is highly different between males and females of C. felis felis; however, this characteristic may be unclear for separating males of the two species.Consequently, in some studies, males of C. felis felis have been incorrectly diagnosed as C. canis.
Morphological variations among fleas do not always result from interbreeding between species (TIPTON; MACHADO-ALISSON, 1972;AMIN;SEWELL, 1977;AMIN et al., 1974;LINARDI, 1984), but they must be used carefully for taxonomic purposes.Since the chaetotaxies of the hind tibia and LMA showed significant intraspecific variations, these two characteristics must be cautiously used for interspecific diagnoses.Sometimes, specimens exhibiting variations have been improperly treated as hybrids, in spite of the nonexistence of the two species in the same municipality or region.
It is concluded that separation of the two species of Ctenocephalides must be done while considering all characteristics.Data on hosts, geographical distribution and prevalence of infestation may support identification of the species.

Figure 1 .
Figure 1.Female of C. canis.A. shape of the head; B. length of the first spine of the genal comb; C. number of bristles on the lateral metanotal area (LMA).

Figure 2 .
Figure 2. Female of C. f. felis.A'. shape of the head; B'. length of the first spine of the genal comb; C'. number of bristles on the lateral metanotal area (LMA).

Figure 3 .
Figure 3. Female of C. canis.Two short stout bristles in the interval between the postmedian and apical long bristles of the dorsal margin of the hind tibia.

Figure 4 .
Figure 4. Female of C. f. felis.One short stout bristle in the interval between the postmedian and apical long bristles of the dorsal margin of the hind tibia.

Figure 5 .
Figure 5. Male of C. canis.Shape of the manubrium of the clasper.

Figure 6 .
Figure 6.Male of C. f. felis.Shape of the manubrium of the clasper.

Figure 7 .
Figure 7. Female of C. f. felis presenting variations in the number of bristles on the lateral metanotal area (LMA): three on the left side, two on the right side.

Figure 8 .
Figure 8. Female of C. f. felis.Variation of the chaetotaxy of the hind tibia.

Table 1 .
Morphological differences between C. felis felis and C. canis.

Table 2 .
Species of hosts for Ctenocephalides spp.recorded in Brazil.

Table 3 .
Endosymbionts of Ctenocephalides species encountered in natural and experimental infections in Brazil.

Table 4 .
Prevalence of C. f. felis and C. canis on companion animals raised in different regions of Brazil.

Table 5 .
Variations in the number (Nº) of bristles on the lateral metanotal area (LMA) and hind tibia of C. f. felis from dogs in the vicinity of Serra do Cipó National Park, state of Minas Gerais.

Table 6 .
Number (Nº)of bristles on the lateral metanotal area (LMA) in 87 specimens of C. f. felis from dogs in the vicinity of Serra do Cipó National Park, state of Minas Gerais.

Table 7 .
Chaetotaxy of the dorsal margin of the hind tibia in 87 specimens of C. f. felis from dogs in the vicinity of Serra do Cipó National Park, Minas Gerais State.