Conventions in Spider Classification
This page outlines the terminology used to identify individual animal species, including spiders, and the animal groups to which they belong.
FUNDAMENTALS OF ANIMAL TAXONOMY
The expression 'animal taxonomy' has traditionally meant the classification of animals into groups based on differences in their anatomy. By convention the
assignment of scientific names to any animal or group of animals is governed by the rules laid down by the International Committee for Zoological Nomenclature.
The most basic division of animals is into phyla, each Phylum being distinguished by just a few fundamental anatomical features. For example, animals such as
fish, reptiles and mammals that have an internal backbone enclosing a long tube of nervous tissue belong to the Phylum Chordata. It can be argued that
the most successful and diversified of all the animal phyla is the Phylum Arthropoda. The name 'arthro-poda' indicates that several segments linked by flexible
joints are found on each leg of this kind of animal. Among the arthropods three subgroups (technically known as Classes) that have been particularly
successful over the millenia are the Insecta, Arachnida and Crustacea with three, four and five pairs of walking legs respectively. Spiders belong to the
Class Arachnida, which is further subdivided into several Orders, notably the Araneae (true spiders) but also the mites, ticks, scorpions, and a number of less
significant arachnids. Details of these spider 'cousins' are summarized on a separate page of this website.
The Order Araneae is made up of a number of Families (with names finishing with '-idae'), each once again distinguished by several significant differences in anatomical characteristics. However, the choice of characters used to define a particular family is somewhat arbitrary and therefore occasionally subject to change. Thus one large family may be split to create two or more new, smaller ones or, less frequently, several small families are combined to become a single large one. Alternatively, a family may be subdivided into two or more subfamilies (with names finishing with '-inae') if the total collection of species within the family seem to divide logically into several distinctly different groups. It is normal practice to derive the family name from the generic name of the first spider of that type to be described in a scientific publication. For example, species of the Araneus, Theridion and Thomisus genera that were first described many years ago were the source of the Family names Araneidae, Theridiidae, and Thomisidae.
Unfortunately, misidentifications by spider taxonomists of the 19th and early 20th centuries have occasionally meant that some existing family names have had to be changed. A good example of this is the Family Sparassidae for which the original family name was derived from the genus Sparassus. This family name was briefly changed to Heteropodidae because it seemed that the first member of this family to be described was actually a Heteropoda species but later reviews resulted in restoration of the original name. Curiously, the generic name Sparassus is not used for any presently accepted member of the Family Sparassidae.
In general, only a relatively small number of characters are used to define each spider family. For example, the fundamental distinguishing feature of all members of
Family Salticidae is a pair of eyes that are much larger than the other eyes and that are aimed forwards like the headlights of a motor vehicle. Similarly, all of the
Oxyopidae have strong leg
spines that are oriented at right angles to the leg surfaces, the Araneidae
typically build insect-trapping orb webs, and the Actinopodidae have a short, broad head region which declines dramatically to the level of the rest of
However, assigning many individual spider species to particular families using only a small number of characteristics has proved problematical because a number of 'exceptions to the rule' have been found. For example, at least some members of the Deinopidae and the Lycosidae have a pair or large, forward-pointing eyes reasonably similar to those on the Salticidae and members of the Family Araneidae are not the only kinds of spiders that build webs that are more-or-less orb-shaped. At least some Agelenidae and Stiphidiidae also do this. Conversely, Araneidae such as Dolophones and Poltys species use camouflage rather than a web to catch their prey although in most other respects they do deserve to be called araneids.
Up until about 1985 spider enthusiasts, including many people with enough knowledge of spider taxonomy to publish books and field guides about spiders, seemed willing to follow the principles behind the popular adage that "If it looks like a dog, barks like a dog, and answers to the name Fido then it probably is a dog!" As described in more detail below, the taxonomists exploring Australia in those early years tended to 'lump' into a single family all species that looked anatomically very similar to each other. More recently, there has been a strong move towards a 'splitter' mentality which favours the creation of more families, genera and species by taking into account a greater number of physical and behavioural characters of individual spider species. Thus some former members of the Family Araneidae are now located in the more recently created Tetragnathidae and Arkyidae.
A wise choice of characters to be used to subdivide an existing spider family into two or more new families is not easily made and proposed relocations of this kind frequently lead to disagreement among spider taxonomists. For this reason some Australian spider genera have been moved to a different family only to be subsequently returned to the original one. An example of this is the leaf-curling spider Phonognatha which was initially placed in the Araneidae then transferred to the Tetragnathidae and later returned to the Araneidae.
One of the ways in which taxonomists have resolved the difficulties associated with dividing a group of rather similar species into several families has been to create a super-family name for the entire group. Thus the Lycosoidea includes spiders which have substantial anatomical similarities with Lycosa species but are not similar enough to be given the Lycosa generic name or even to be placed in the Family Lycosidae. In Australia the lycosoid families include the Lycosidae, Pisauridae, Ctenidae, Zoridae and Stiphidiidae, the species of all of these having a reasonably similar overall appearance.
The creation of 'off-shoot' spider families from existing ones often means the members of the new families are not very different in overall appearance from those in the 'parent' family and this makes their classification more difficult except when performed by an expert arachnologist. For a long time few people could distinguish between the flat-bodied spiders Hemicloea and Rebilus (now Morebilus) and this confusion was understandable because Hemicloea (Family Gnaphosidae) and Morebilus (Family Trochanteriidae) have almost the same overall appearance and are only distinguished by careful examination of structures such as the spinnerets which cannot be seen easily without the aid of a stereo microscope.
In animal taxonomy perhaps the only term used that does not involve a degree of conjecture and somewhat subjective assignment is the species name. Taxonomic convention (at least for printed publications) states that both the generic and the species name of each spider should be written in italics, the genus starting with an upper case letter and the species name all lower case letters. Note: This convention has been followed on this page but elsewhere on this website the italics have generally been omitted to avoid confusion, bearing in mind that this website is primarily intended for use by non-experts rather than scientists. Trivial or common names for particular spiders are not italicized even in scientific publications. Because the names assigned to many species have changed over the last century or so the scientific name is correctly followed by the name of the person who created it along with the date on which his description of it was first published. Thus, the cosmopolitan daddy-long-legs spider is Pholcus phalangioides Fuesslin, 1775. However, the thomisid Sidymella trapezia (L. Koch, 1873) has Koch's name in parenthesis because he originally named this species Stephanopis trapezia and it was only moved to the Sidymella genus at a later date.
Two animals are the same species only if they are capable of mating and producing viable and fertile offspring. While this may not always be achievable in practice (as when a male Chihuahua attempts to mate with a female Doberman) it remains the accepted definition of a species. Of course, evolution is an on-going process so we have to accept that some individuals of a single species will eventually undergo genetic and anatomical changes that make them no longer able to breed successfully with unchanged specimens of that original species. This process is normally a gradual one so there may be a period in which interbreeding between the old group and the new one is at least partially successful.
However, in the case of many kinds of spiders the ability to interbreed successfully is not always easy to test. One reason for this is that for some species only the male or the female has ever been collected and formally described. It is true that for certain species the presence of the male in constant proximity to the female and the subsequent, if somewhat reluctant, willingness of the female to allow him to mate is convincing evidence that both sexes are the same species even if strong sexual dimorphism between the male and female suggests otherwise.
Because the ability to reproduce successfully is the fundamental criterion for species recognition it has always been considered essential when describing a new species to clearly state and illustrate the appearance of the female epigynum and the genital structures underneath it as well as the shapes of the terminal segments of the male palps. Of course, the overall appearance of the spider's body, legs, and spinnerets are also taxonomically useful but these are probably of greater value for assignment of a spider to a family rather than for deciding the appropriate genus or species name for it.
While humans may sometimes fail to recognize that a male and a female spider with obvious differences in appearance are actually the same species this is mostly not a problem for the spiders themselves. It is probable that this is at least partly because of specific pheromones, usually released by the female, and if this is the case then these pheromones might also have taxonomic usefulness. Unfortunately, the chemical nature of only a very small number of spider pheromones has been determined so far. In addition, the amounts of pheromone released by each spider are too small to be easily collected and analysed and release probably only occurs in certain circumstances such as the arrival of the 'breeding season', the achievement of maturity by the male and female, and the female's mating 'history'. The benefits and limitations of spider identification by pheromone analysis have been well reviewed in the following paper:
The comment in Gaskett's paper that "sex pheromones are not necessarily species specific, but may still assist with species recognition" is worth noting. It probably is fair comment to state that if a male spider clearly has been attracted by a pheromone to a female the two probably are the same species but if there is no evidence of sexual attraction of the male by the female this is not conclusive evidence that the two are different species.
For many Australian spiders one of the two sexes can be very difficult to find in the field. Male spiders only manifest the appearance of the adult male at maturity or perhaps at the penultimate juvenile stage. This may also be the only time in their life cycle when the female's pheromones stimulate them to attempt to mate. Conversely, the females of some species hide in retreats so well camouflaged that they are almost impossible to find except by chance. This is especially true for many mygalomorph spiders where it is the role of the males to wander above ground in search of females which spend almost their entire lives in their burrows.
The difficulty in proving that certain males and females are actually the same species has sometimes led to the two being assigned different species names. Recently, Brodie Foster, a salticid enthusiast living on the Sunshine Coast of Queensland, noted that whenever he found plenty of Opisthoncus parcedentatus females in a particular location there were also numerous Opisthoncus mordax males in that same area. Since he could find no convincing evidence that males of O. parcedentatus and females of O. mordax had ever been described, Brodie suggested that the male presently known as the O. mordax male is actually the male of O. parcedentatus. He even watched a spider that had the markings of an immature female O. parcedentatus moult to take on the appearance of an adult Opisthoncus mordax male.
The term genus can be defined as "a taxonomic category within a family and composed of a group of species exhibiting very similar characteristics". Once again there is an arbitrary aspect to the sorting of the spiders in a single family into multiple genera. Two spiders that have almost the same anatomical and behavioural characteristics but that do not interbreed will normally be placed in the same genus but with different species names. This system leads to inevitable disagreement among spider taxonomists as to the appropriate generic name for some species. Many amateur spider enthusiasts may at times be surprised and frustrated that a spider for which they thought they knew the scientific name is now presented in a newer book or website under a different name. This can even be true of the family a spider is placed in, new families occasionally being erected using some members of an existing family (e.g. the Arkyidae were once part of the Araneidae) or, less often, a complete family is subsumed into an existing family (e.g. it is now being proposed that the Family Zoridae should no longer exist, its members being placed in the Family Miturgidae).
Why do these name changes occur? Well, one major reason for this is that almost all of the taxonomists who first collected and described spiders in Australia more than a century ago were from European countries such as Germany, France and Britain so, not surprisingly, these people tended to assign European generic names to Australian species that looked very similar to the Northern Hemisphere spiders with which they were already familiar. For example, they described many new Australian Araneus and Theridion species, almost all of which are now likely to be given new generic names as taxonomists revise the families to which they belong. In addition, new techniques such as high-resolution imaging and DNA comparisions, which were not available to those early arachnologists, are continuing to show that many of the original names assigned are inappropriate.
However, opinions as to the extent to which this renaming should occur vary according to whether the taxonomists involved are 'lumpers' (who prefer to have fewer families, genera, and species) or 'splitters' (who believe there are valid reasons for subdividing existing groups). In the 21st century the latter group predominate as methodology for distinguishing between closely related species improves. In recent years this methodology has also revealed many 'cryptic' species complexes, these consisting of two or more species that are so nearly the same in appearance that they are initially assumed to be a single species although they actually cannot interbreed. A good example of this is the so-called 'Araneus acuminatus group'. Similarly, in 1980 most people thought there was only one funnel-web spider and one white-tailed spider but we now recognize three Australian funnel-web genera (and 35 different species) and nearly 200 species of Lamponidae (the white-tailed spider family).
At the species level European spiders and very close relatives on other continents are sometimes distinguished by the use of subspecies names. For example, the Mediterranean black widow spider was named Latrodectus mactans tredecimguttatus, the North American one was Latrodectus mactans mactans and the Australian redback spider became Latrodectus mactans hasselti. However, the middle name for these three spiders has recently been deleted and each subspecies is now considered to be a separate species although, as can be seen from the following illustration, the pattern of markings on the upper abdominal surfaces of these Latrodectus species is not a reliable means of distinguishing them. Curiously, it appears that data verifying that the three original Latrodectus subspecies are unable to interbreed has never been published.
Because of the great distances involved and limited communication systems when arachnologists first wandered around Australia it was inevitable that two or more of them independently collected the same species but gave it a different scientific name. As a consequence of this today's taxonomists frequently have to make extensive searches of museum collections looking for the original preserved specimen and any apparently similar ones that were collected at a later date and either given a different name or given the same name when they are actually different species. To avoid confusion authors of revision papers normally list as synonyms all alternative names that a particular species has been given. It is a curious fact that the first Australian funnel-web spider to be described was Hadronyche cerberea, the more famous Atrax robustus only being named five years later. The females of funnel-web species look almost identical and in 1980 all funnel-webs were considered to be Atrax species. Since that time differences in the males have led to the current recognition of three different funnel-web genera.
Unfortunately, the discovery and naming of all of the spiders of Australia is far from complete and many revisions of earlier classifications are overdue to be carried out. In consequence, virtually all comprehensive books and websites that name Australian spiders contain misidentifications, many of which are because of recent changes to generic and/or species names. Taxonomic revisions have also tended to involve the creation of new generic and sometimes even new family names for Australian spiders as well as to the subdivision of what was once thought to be a single species into several different species. Recognizing these new species is often difficult or even impossible without extensive anatomical knowledge and access to sophisticated lab equipment. Only in instances where a particular species has very distinctive visible features can it be confidently identified if only a single photo of it is available. The absence of any information as to the habitat and locality in which the spider was found also make its identification more difficult since each continent, and each district within a large land mass, has its own unique collection of spider species. For this reason sets of photos of American or European spiders are only of very limited use for identification of Australian species.
SPIDER IDENTIFICATION BY DNA TECHNOLOGIES
A major reason for the development of DNA-based methods for identifying humans and other living creatures was the observation that the transplantation of hearts and other organs between unrelated humans almost always led to immunological rejection of the transplanted organs by the recipient's body. To overcome this problem and also to find ways to better manage inherited diseases researchers set out to identify the genes responsible for the development of each body trait. It was hoped that methods could be devised to match organ donors and recipients, thereby minimizing rejection problems, and also to be able to predict who would develop an inherited disease state. From this early work came the Human Genome Project, genetically modified plant, animal and microbial species, and the technology to identify individual humans and animal species by determining the gene contents of the DNA of their tissue cells.
DNA 'fingerprinting' involves the extraction of DNA (which takes the form of a long double strand of individual nucleotide units called base pairs) from a sample of tissue cells and its fractionation into fragments that will spread out along an electrophoresis gel strip in a pattern reminiscent of the barcode on a grocery item. DNA fingerprinting now supplements conventional fingerprinting for forensic testing but its use has been hampered by the fact that it is a much slower and more expensive technology. For those interested in learning more about the procedures involved in preparing DNA barcodes for identification purposes the following websites may be worth visiting:
The more complex animal groups, including spiders, all have far too many genes for it to be justifiable to create a fingerprint of the entire DNA content of each cell just for identification purposes. Fingerprinting the entire DNA contents of a species would be a very expensive and time-consuming process and would usually be pointless in that most of the genes in a particular kind of animal will be the same anyway. However, it is now not too difficult to use enzymes to dissect DNA into smaller pieces and to select and replicate just a few key fragments.
The only cell components that contain DNA are the nucleus and the cytoplasmic mitochondria, the latter having only a small loop of DNA (at least in humans) whereas the chromosomes in the cell nucleus have very long coiled DNA strands. Early in the first decade of the present century research papers began to be published in which it was being claimed that the matching of DNA from the mitochondria of spider tissue cells rather than from chromosomes was easier to perform but virtually as reliable taxonomically.
Mitochondria are small oval organelles with only a few functions, one of the most important of which is the supply of energy for the cell's many metabolic processes. Particular attention has been focussed on the mitochondrial DNA gene that codes for the enzyme cytochrome C oxidase subunit 1 (usually written COI). This contains only 648 nucleotide base pairs and so can be barcoded relatively quickly. The COI barcode has been found to vary substantially from species to species but only slightly within a single species and for this reason has now been determined for a wide variety of animals.
Unfortunately, present evidence indicates that the use of the mitochondrial COI base pair pattern as the sole criterion for species identification is not completely without risk because of unsatisfactory lab procedures and also because it has now been shown that occasionally two apparently different species share the same COI barcode or, more frequently, individuals that are otherwise unquestionably the same species have different COI patterns. Multiple COI barcodes among the individuals of a single species might be the result of mutation, accidental contamination of the sample with DNA from another source, or infection of the host animal's cells by microorganisms. It is significant that at least for humans mitochondrial DNA seems to mutate 10 - 100 times more frequently than nuclear DNA though why this should be so is unknown. The diversity of human disease states that can be the result of mitochondrial gene mutations is well illustrated in the website of the Neuromuscular Disease Centre at St. Louis, Missouri:
The benefits and limitations of DNA barcoding for animal species are well laid out in the following Wikipedia website page:
It is clear that the idea that species identification by COI matching alone is the ideal way to identify species in the twenty-first century is a long way from being accepted by all eminent taxonomists. On the other hand, the matching of one or preferably a few gene segments is a powerful method for distinguishing between two species that appear to be almost identical on the basis of all relevant anatomical characters. This is particularly true for cryptic species belonging to a single genus and hopefully will reduce the need for taxonomists to use expressions such as the 'Araneus acuminatus group' in their publications. A recent paper by Framenau et al that illustrates the use of the COI gene for identifying and separating spider species within a cryptic generic group can be accessed from the following URL:
THE CLADE CONCEPT
Progressively more anatomical and behavioural information is employed as a taxonomist moves from Phylum to Class then to Order, Family, Genus and Species. This is a logical approach to take since the best way to create subsets within a large group is to take into account progressively larger numbers of distinguishing characters and the concept of evolution implies that time and altered environmental circumstances cause progressive small changes in the genome of each species until eventually these changes are so great that the descendents of the original species have diverged into two or more new species that can no longer interbreed successfully.
As illustrated in the following graphic, a clade is a branching diagram that shows the probable way that a group of families, genera or species have evolved from a common ancestor. In other words, it is a branch of the so-called 'tree of life' which attempts to display how one or a few very primitive animals evolved over a long period of time into the great diversity of animal life that exists today.
The range of spiders included in a clade diagram can be as large or as limited as suits the taxonomist who is attempting to create it. Clade diagrams may be drawn up to show how new families have been formed as off-shoots of existing ones and to suggest the likely order in which new genera or species have evolved from more primitive ones within a single family. On this basis a clade can be like a major branch on a tree trunk or just a group of fine branches at the growing tip of that large branch.
It is important to understand that a clade is constructed by statistically analysing a large number of anatomical and perhaps behavioural traits that are characteristics of that group of animals and placing closest to each other on the diagram those groups that seem nearest to being identical. The logic behind this process is that evolution is a gradual process so two species that are almost the same will probably have evolved from a parent species more recently than two species that have far fewer matching characteristics. The increasing availability of sophisticated taxonomic tools such as scanning electron microscopes and DNA fingerprinting will undoubtedly make both species recognition and clade construction more accurate in the future than it has been in the past.
Email Ron Atkinson for more information. Last updated 13 May 2016.