A typical mammalian spermatozoon consists of a head partly covered by an acrosome, a neck and a flagellar-like tail. The head of the mammalian spermatozoon is ovate, ensiform or falciform and dorsoventrally flattened. The neck typically consists of the connecting piece and the centriole . The mammalian sperm tail contains an axonemal complex of microtubules and a further nine outer dense fibers to complete the 9+9+2 pattern . In the midpiece of the mammalian spermatozoon, the axoneme and outer dense fibers are enclosed by a long sheath of mitochondria. The mitochondria itself are elongated and arranged around the core of the sperm tail in a helical fashion. The number of mitochondrial gyres varies between mammalian species, with the relatively short midpiece of the human consisting of about 15 gyres, whereas the exceptionally long midpiece of several rodent species contain as many as 300 gyres . In the principal piece of the flagellum, the axonemal-outer dense fiber complex is surrounded by the fibrous sheath, which is divided into several transverse ribs along the length of the principal piece .
The spermatozoa of naked mole-rats in this study deviate markedly from that of virtually all mammals. The sperm head surface is extremely irregular and often form small or large lobes with a high percentage of either dispersed chromatin or so-called nuclear vacuoles. The lobed nucleus in particular appears to be degenerate compared to that of most mammals. Together these morphological attributes would result in most naked mole-rat spermatozoa being classified as "abnormal". Importantly, these attributes are not considered to be a major consequence of inbreeding as the individuals used in this study originate from colonies in a captive population that include both inbred and outbred pedigrees and a low mean level of inbreeding (F = 0.163) . The high inbreeding coefficient reported in a previous study (F = 0.45) among four wild-caught colonies of naked mole-rats in Kenya , can be due to the fact that three of these colonies were collected within 5 km of each other. New colonies of naked mole-rats are usually formed by fissioning and thus neighbouring colonies could have a recent common maternal ancestor [23, 29].
The neck of the naked mole-rat spermatozoon contains a poorly developed capitulum which gives rise to five banded columns. In most mammals and particularly in rodents, however, the capitulum represents a large, solid and well developed structure . The well defined midpiece of most mammalian species, particularly in terms of the highly organized helical/non-helical arrangement of mitochondria, is replaced in the naked mole-rat by a small and generally disorganized midpiece. The midpiece length is the shortest of all mammals so far recorded (± 1 μm)  and the total number of mitochondria (± 7) is also among the lowest for any mammalian species . Furthermore, the mitochondria are randomly dispersed and their form varies even within the same sperm midpiece. Accordingly, the midpiece of naked mole-rat spermatozoa appears to show various degenerate features.
The greatest deviation from the mammalian pattern in the naked mole-rat spermatozoon is the structure of the principal piece of the sperm tail. In this species, there is no apparent difference found in the size of the nine outer dense fibers surrounding the axoneme. However, in many mammalian species the outer dense fibers numbered 1, 5 and 6 are distinctly larger than the other six fibers and some species also have a larger fiber in position 9 . Although the 9+9+2 pattern persists in naked mole-rat spermatozoa, there is no fibrous sheath present. One of the main suggested functions for the fibrous sheath is to provide structural support/strength to the tail beating rapidly in a viscous medium as encountered in the female reproductive tract [26, 31]. Structurally these deviations in the principal piece thus represent further simplified and possibly degenerative features of the naked mole-rat spermatozoon.
Sperm structure has been extensively used as a tool to assist in both taxonomic and phylogenetic studies and more recently as an indicator of relative sperm competition. For example, van der Horst et al.  showed that acrosome structure and shape can be used to distinguish among four very closely related ferret species. Breed [33, 34] has furthermore shown that sperm head structure is related to phylogenetic relationships in rodents in addition to their phylogenetic derivation (primitive versus advanced structures). However, despite the fact that certain species' spermatozoa may be derived or have become more specialized or simplified, one seldom encounters the situation where there is such a large variability in sperm form within a given species as observed in the naked mole-rat. Human sperm provides a rare example of sperm polymorphism and in human clinical spermatology, any deviation in sperm structure from normal is defined as abnormal according to the so called Strict Criteria . Thus, in this study we revealed that, similar to humans, naked mole-rat spermatozoa have a high degree of polymorphism.
An important question that emerges is: which of these "polymorphic" spermatozoa are normal or abnormal and how does sperm morphology affect their ability to fertilize an oocyte? During standard semen analysis procedures, the normality of sperm morphology is an important characteristic in determining the fertilizing potential of spermatozoa [35, 36]. In many mammalian species (natural populations) a relatively high percentage of spermatozoa in the ejaculate are morphologically normal (> 80%) . In most of these species the level of sperm competition is high and it can be assumed that there is strong selection pressure to produce a high percentage of spermatozoa that are structurally and functionally normal . The end result is that most spermatozoa have an almost equal chance of fertilizing an oocyte.
Previous studies which mentioned the existence of variation in male fertility of some mammalian species, still reported a relatively high percentage of morphologically normal sperm, e.g. 77% (range 12-97%) in natural populations of red deer  and 76% (range 6-91%) in adult dogs . Even the endangered black-footed ferret (Mustela nigripes), which is exposed to a high degree of inbreeding, had 68% normal spermatozoa in the breeding season . Interestingly, in humans who typically have a low risk of sperm competition, males with more than 15% normal spermatozoa is regarded as fertile according to Strict Criteria  and the lower reference limit for normal forms is 4% . Preliminary results from our laboratory have shown that the naked mole-rat has only few normal spermatozoa (± 5-15%, data not shown). Thus, most of the polymorphic spermatozoa in humans and naked mole-rats are apparently abnormal and accordingly not variations of normal spermatozoa.
Consequently, sperm competition would appear to be extremely unlikely in naked mole-rats and there seems to be no need to produce perfectly formed and highly motile sperm. Parker  emphasized this principle by mentioning that the production of high quality, error free spermatozoa is costly and that there will be selection against it if the costs are not equal to or outweighed by the benefits (fertilizing the oocyte). Thus, in the absence of sperm competition, there may be little benefit in investing energy on the quality of sperm production . However, when there is a high risk of sperm competition, every sperm counts and selection will favour the production of high quality spermatozoa . Further evidence for the absence of sperm competition in the naked mole-rat is entrenched in its sperm structure. Both the short midpiece and short tail (± 28 μm) of naked mole-rat spermatozoa is typical of mammals with a low risk of sperm competition [8, 9, 45]. The possible effect of a lack of sperm competition on the size and structure of spermatozoa is indirectly emphasized by Lijfeld et al.  who reported that an increased risk of sperm competition selects for longer spermatozoa and reduces between-male and within-male variation in the sperm length in passerine birds. However, another factor contributing to the small size of naked mole-rat spermatozoa could be this species' lower metabolic rate . Recent studies on the effect of metabolic rate on sperm size [48, 49] have shown that there is a positive relationship between the mass-specific metabolic rate and the total sperm length of both eutherian and marsupial mammals and that species with a lower mass-specific metabolic rate produce uniformly small spermatozoa [48, 49].
Despite the fact that few of the naked mole-rat spermatozoa are structurally and physiologically normal (e.g. motile), the breeding males in this study were all producing healthy litters of pups prior to their removal. This suggests that they are capable of producing sufficient normal spermatozoa available to fertilize multiple oocytes. The relatively high average sperm concentration found in the current study was probably due to the fact that a very high sperm concentration (100 × 106/ml) was measured in only one male and therefore skewed the data. However, for most males the sperm concentration varied between 5-50 × 106/ml and the lower limit of the current study is comparable to the 1.8-8.6 × 106 spermatozoa in one half of naked mole-rat reproductive tract previously reported by Faulkes et al. . This low average concentration of spermatozoa in the naked mole-rat could be another effect of the absence of sperm competition and is consistent with the theory that an increase in sperm competition will increase the number of sperm produced by a male . An extreme case of this phenomena is found in the yellow seahorse (Hippocampus kuda), a species which also lacks sperm competition, where the testes only contain about 300 spermatozoa  and results in a sperm:egg ratio comparable to that of the social insects .
The degenerative structural state of both the midpiece and the tail could possibly explain the poor motility of naked mole-rat spermatozoa. The kinematic parameters clearly showed that there was both low percentage motility as well as sluggish moving spermatozoa. The low percentage motile spermatozoa recorded in the current study (7.3%) were similar to the 5% motile spermatozoa reported by TB Hildebrandt [personal communication]. Although Faulkes et al.  also found relatively low percentage motile spermatozoa in naked mole-rats (< 50%), they reported a significantly lower sperm concentration and lower percentage motile spermatozoa in subordinates relative to breeders, which were not evident from the current study. The spermatozoa in the current study were swimming at an average curvilinear velocity of 35 μm/s, which may well be the lowest recorded for any mammalian species. In other social mole-rats of the same family (Bathyergidae), the average sperm velocity is 148 μm/s and a high percentage of sperm motility is evident . Even humans, who have a high percentage of abnormal spermatozoa, typically have more than 60% motile sperm and they swim with an average velocity of about 90-120 μm/s [18, 54, 55]. The slow swimming speed of naked mole-rat spermatozoa could thus be the result of both the short tail, which beats with a lower force, and the small midpiece with few mitochondria, which may generate less energy for motility.
Another aspect that requires attention is how much simplification or degeneration is present in naked mole-rat spermatozoa? Part of the answer may be found by looking at sperm structure in the monotremes such as the platypus. Here typical mammalian sperm features are maintained and the fibrous sheath of the principal piece of the tail is well developed . In contrast, marsupial spermatozoa seem to share sperm characteristics with the sauropsids rather than mammals and therefore reflect a more primitive condition. However, even in this instance the fibrous sheath is a typical feature of the principal piece of the tail. The absence of this feature in naked mole-rat spermatozoa when compared to the primitive mammals accordingly supports the view that this is a degenerative feature in naked mole-rats and not a primitive or simplified feature. Naked mole-rat spermatozoa seem to be derived from ancestral rodent sperm with a hooked acrosome. Breed  concluded in his study on rodent spermatozoa that, "as the hook-shaped sperm head and long sperm tail occur across the muroid subfamilies, as well as in the heteromyid rodents, it is likely to be the ancestral condition within each of the subfamilies with the various forms of non-hooked sperm heads, that are sometimes associated with short tails, being highly derived states". The low number and disorganized nature of the mitochondria in the midpiece of naked mole-rat spermatozoa also appears to show degenerate features rather than simplification. When a spermatozoon has simplified, there is usually great order in terms of its organization (e.g. teleost sperm) and contrasts sharply with the situation in naked mole-rats.
We evoke the term 'degenerative orthogenesis'  to describe the degenerate appearance and poor motility of naked mole rat spermatozoa. According to Gould  (also  and ), Wilhelm Haacke devised the word "orthogenesis" which means "straight (line) generation" and subsequently "orthogenesis denotes the claim that evolution proceeds along defined and restricted pathways" . In this context Gould  based his interpretation on dissecting the work of some eminent evolutionists of their time [58, 60–63]. While it is considered as a formalist theory standing against the central Darwinian principle, it has been interpreted in a broader context by many [64, 65]. It is particularly Gould  that supports a modern use of processes/concepts generally described as saltations (discontinuous evolution, constraints) and channels (internally generated pathways) and includes orthogenesis to understand evolutionary change within the Darwinian framework. These notions above represent two sides of Gould's conviction that the internal structure of an organism can set and constrain the pathways of change .
Degenerate animals often have a simpler anatomy than primitive and non-degenerate animals, such as in Lepas . De Villiers  furthermore emphasized that it is not only the individual animal of a species that is sensitive to stimuli from the environment, but also the embryo and larvae. For example, de Villiers  indicated that, in vertebrate embryos, there appears to be a delayed development of certain openings and tracts due to the pressure of assimilated yolk inherited from their ancestors. Accordingly gametes would also be exposed and respond to various stimuli and undergo changes. However, the authors, in agreement with de Villiers , do not suggest that all these changes are palingenetic but rather kenogenetic. Morphological degeneration is not a new concept and Eimer  referred to this as an environmental impetus of a balance between internal and external forces. However, it was viewed in a narrow formalist context which was difficult to analyze scientifically  and therefore required interpretation in a broader framework.
Most structures in naked mole-rat spermatozoa clearly became degenerate, such as components of the head, midpiece, neck and rest of the flagellum. It is important to draw a clear distinction between sperm degenerative features due to inbreeding and those due to the absence of sperm competition. Pure inbreeding degeneration in sperm structure may partly include features such as sperm DNA fragmentation  and sperm morphology abnormalities (abnormal size and shape of the head, midpiece and tail) [67–70]. Degenerative changes due to virtually no sperm competition, however, involve a vast simplification in features, for example the absence of the fibrous sheath in the principal piece of the tail (a fundamental mammalian sperm structural feature ), an abbreviated midpiece with few simplified mitochondria and a poorly developed capitulum in the sperm neck. To our knowledge, this is the first study to describe the presence of such "degenerative features" in a mammalian spermatozoon.
Hence, in naked mole-rat spermatozoa it appears both inbreeding and the absence of sperm competition may have contributed to abnormal sperm features but that the degenerative features mentioned above represent very specific absence or modification of structures such as the midpiece and tail. It is possible that natural selection forces operated, but that simplification in sperm structure was primarily driven by the lack of sperm competition. This apparent absence of sperm competition was followed by a morphological degeneration of sperm structures, representing a process of degenerative orthogenesis, and is largely based on their reaction to the internal environment. There does not appear to be any advantage or adaptation in this degeneration of sperm structures and the spermatozoa simplified or degenerated to such an extent that it is on a path of no return. In this investigation our interpretation is in line with Gould  who considers these older formalist concepts in a broader context in assisting to understand the theoretical base of evolution within a Darwinian framework. Furthermore, our research presents a unique finding that evolutionary processes such as degenerative orthogenesis may occur right up to the cellular level and not only in the individual or embryo as was previously shown.