The Dormouse Hollow

Dormouse Conservation

A Review of the fat dormouse (Glis glis) in Britain

The following text is taken from:

Morris, P. A.: A review of the fat domouse (Glis glis) in Britain, Nat. Croat., Vol. 6, No 2., 163-176, Zagreb 1997.
ISSN 1330-0520 / UDK 598.323+591.5(410.1)
Nat. Croat., Vol. 6, No. 2., 163-176, Zagreb, 1998. Hrvatski prirodoslovni muzej, Demetrova 1, Zagreb, Croatia



Contact: Dr. Pat Morris: School of Biological Sciences, Royal Holloway, University of London, Egham, Surrey TW20 OEX, U.K.

The fat dormouse (Glis glis) was introduced to Britain in 1902. The species is still confined to a relatively restricted area. Its legal status is confused, being both legally protected, but at the same time an alien species and significant pest. Licences may be granted to trap Glis, but they may not then be released into the wild because they are not a native species in Britain. The total British population is unlikely to be less than 10,000. Nobody seems to be regularly monitoring Glis in Britain, but some small studies have now begun aimed at understanding more about population density, movements and potential ways of controlling this animal.

Key words: Glis glis, Britain, ecology, monitoring


Two members of the Gliridae occur in Britain. The native species is the common or hazel dormouse (Muscardinus avellanarius) This is fully protected under the Wildlife and Countryside Act, 1981 (Schedule 5), and its historic decline in numbers and range have now been recognised and perhaps halted (BRIGHT & MORRIS, 1996a; BRIGHT & MORRIS, 1996b; BRIGHT, MORRIS & MITCHELL-JONES, 1996). The second species is the fat or edible dormouse (Glis glis). It was introduced to Britain from the Continent in 1902 by Walter (later, Lord) Rothschild. Despite being numerous where it occurs, it is still very localised in its distribution, and therefore not considered to be nationally important. Nevertheless, public interest is high and a BBC television film "The Incredible Edible Dormouse" was seen by over 6 million viewers when it was first broadcast nationally in 1990, with high audience approval ratings. However, little is known of the basic ecology of Glis in its English home; this paper provides a review.


According to FLOWER (1929) Glis glis was introduced to Tring Park (Hertfordshire), about 100km north west of London, in 1902. This date has been accepted by most later authors except HARRISON MATTHEWS (1952), who said that introduction occurred about 1890'. FLOWER (1929) also alleged that they were released on or about 4th February. This seems an unlikely time to release a hibernating species. The number of animals liberated remains unknown and their original source is also in doubt. According to FLOWER (1929), they probably came from Germany or Switzerland', a statement repeated by authoritative authors such as FITTER (1959) and LEVER (1977). However, according to JONES-WALTERS (1990), the founder population originated in Hungary. This information was obtained from Mr D. Mead, a pest control officer working for Dacorum local authority, the grandson of the man who obtained the original stock. He was certain that the animals had come from Hungary (L. JONES-WALTERS, pers comm). Shortly after release, the animals were blamed for damage to crops and efforts were made to exterminate them (FITTER, 1959). By the mid 1920s, the dormice were numerous around Tring Park and first reports of damage and nuisance indoors were noted (THOMPSON, 1953). Domestic intrusions were well documented from 1935 onwards, in particular gnawing of electric wires and cases of drowning in household water storage cisterns and lavatory bowls (6 records involving 20 animals; THOMPSON, 1953). Nearly 600 were trapped in houses between 1943 and 1961 in the Amersham area alone (COWDY, 1966).


Previous distribution surveys (THOMPSON, 1953; JONES-WALTERS, 1990) indicated that Glis was confined to woodlands in the Chilterns area, within about 35km of its point of introduction. There are a few outlying records as far west as the Bledlow Ridge, east to at least Potters Bar and south to High Wycombe. Since the survey by THOMPSON (1953), the range of the species has enlarged (JONES-WALTERS pers. comm.), but not by very much. These studies were supplemented in 1995 by a questionnaire enquiry (P. A. Morris, unpublished) organised by the Mammal Society and The Royal Forestry Society, in collaboration with Buckinghamshire County Council and the Hertfordshire Records Centre. This latest survey, confirms the pattern revealed by others, that since their arrival in Britain, fat dorniice have spread very little after nearly 90 years. However, the 1995 questionnaire survey did add a significant number of 1km squares to the distribution of 15 years earlier, suggesting at least some increase in numbers and distribution in that time. Most of the 1995 records (68%) came from Buckinghamshre and Hertfordshire (30%), with one record from as far as Stevenage, about 30km east of Tring. To the west, the spread of the dormouse seems to be constrained by the open farmland of the Vale of Aylesbury. Expansion in other directions has been slow, probably impeded by increasingly fragmented woodland habitat and substantial barriers such as major roads and urban areas. These obstacles probably account for the paradox that Glis is numerous where it occurs, but population pressure has not resulted in significant expansion of its range. The core of the distribution occupies about 50 square kilometres in the Chiltern hills. Here there are extensive and well-connected areas of high forest (mostly dominated by beech, Fagus sylvatica), which apparently provides an optimum habitat for the species. There are also many smaller areas of plantation forestry, including various conifer species, particularly larch (Larix decidua) and spruce (Picea abies). Half of the Glis recorded from plantations in the 1995 survey were present among conifers and a quarter in beech. The plantations and mature woodlands of the Chilterns closely mimic the high forest habitat favoured by Glis in its natural haunts in Continental Europe. However, the majority of 1995 records (78%) were from houses; only 10% were from woodlands and 11% from gardens. This suggests that Glis in houses may now be a more prevalent problem than in woodlands, where their depredations have hitherto been considered more serious. Enquiries reveal that people troubled by Glis in their houses capture the animals, but do not wish to kill them (for the average householder there is no easy and humane method of doing so). The animals are then released, often far away. By this means, Glis is able to cross what would otherwise be barriers to dispersal, leading to significant range extensions. However, release of live Glis is illegal. Many people know this and are therefore unwilling to supply details. Thus the full extent of translocations is under recorded. Nevertheless, reliable (but anecdotal) evidence suggests that Glis has been released near Oxford, some 45km west of Tring (thereby vaulting the Vale of Aylesbury) and in the New Forest, about 100km to the south west. Both these areas offer excellent habitat for the species. A number of widespread, but isolated reports suggest that Glis has been transported (accidentally or deliberately) considerable distances from its point of introduction. One was recorded from Shropshire in 1941, over 200km to the north west (HARRISON MATTHEWS, 1952) and others have been reported from various English counties (FITTER, 1959). A recent record was from near Windsor (Berkshire) and the Atlas of British Mammals (ARNOLD 1993) shows one record of Glis from Sandy in Bedfordshire, about 45 km north east of Tring, presumably also the result of assisted dispersal. It is only a matter of time before the species and its associated problems become commonplace in other parts of England.

Fig. 1. The distribution of Glis glis in England. The arrow indicates Tring Park, site of the original introduction in 1902. Black symbols cover lkm squares where dormice have been recorded 1980-1995 inclusive. Data from Questionnaire survey, 1995 combined with a previous survey by L. Jones-Walters (pers. comm.)

The pattern of dispersal is obscured by misidentification of grey squirrels (Sciurus carolinensis), a common species whose young superficially resemble Glis and probably account for some of the outlying records, including one from Eastleigh (Hampshire) reported in the 1995 questionnaire survey. Most of the records (77%) obtained from the 1995 questionnaires were of live animals, mostly heard (79%) and nearly half were noticed during the morning (unexpected for a nocturnal animal). The majority of respondents (64%) said that Glis had been present for more than five years, but 11% said that the animals had only recently arrived and, in answer to another question, 34% suggested that the animals had remained for one year or less. This information suggests an established population, probably with continuing expansion. It also suggests that the animals can be present without being noticed regularly because many of these new' sightings were from areas where the animal has been present for many years.


The legal status of this species in Britain is very confused. Under the requirement of the Berne Convention for signatory nations to enact legislative protection for all the Gliridae, Glis glis is listed on Schedule 6 of the Wildlife & Countryside Act, 1981. This identifies those partially protected species which cannot be trapped without a licence. However, as it is an alien species, it is also included on Schedule 9 of the 1981 Act which lists those animals whose release into the wild is forbidden. Thus the status and fate of captured animals is left unresolved, the presumption being that all captures should be killed, apparently in conflict with the principle that they should be protected throughout Europe!

If animals are to be trapped out of doors in defence of crops (forestry or agriculture), the relevant licensing authority is the Ministry of Agriculture Fisheries and Food. Where the animals cause a serious nuisance in domestic property, licences to trap are issued by the Department of the Environment; which also has responsibility to grant licences where release is intended (eg. for research purposes). In all these cases the Nature Conservancy Council for England ('English Nature') must be consulted. Thus three Government agencies are involved with this one mammal and none seems sure what to do about it. The Berne Convention allows control of animals which damage property or crops or pose a danger to health or safety, but where this exemption from the terms of the Convention is invoked, a bi-ennial report has to be made to the Standing Committee of the Convention. This report is supposed to include a scientific assessment of the impact of actions taken, but nobody seems to be monitoring Glis in Britain, nor is it clear which Government Department is responsible for doing so. It appears that the United Kingdom may be falling short of its international obligations in respect of this species.


In the absence of quantified fieldwork it is impossible to do more than guess at the total numbers of Glis in Britain, based upon certain assumptions which may be modified by future studies. A survey by JONES-WALTERS (1990, pers. comm.) indicated presence in 70 lkm squares in the period 1980-88, some of which are heavily wooded. A population estimate may be made, based on 3 assumptions:

1. Glis is present in all these 1km squares, but no more. 2. Within these 70 squares Glis is present over the whole of the total area they represent. 3. Glis occurs at an average density of 1 per hectare. This is based on a small study by HOODLESS & MORRIS (1993) and is also the estimated population density in Czechoslovakia (GAISLER, HOLAS & HOMULKA, 1977), but well below the numbers suggested in some other areas of Europe, where densities are thought to exceed 50 per hectare.

These assumptions result in an estimated total population of 7,000 animals. However, Fig. 1 suggests a core area of about 100 kilometre squares when recent records (from the 1995 questionnaire survey) are added. This would indicate a population of nearer 10,000 animals. Moreover, 75 Glis were reported from one house (THOMPSON, 1953) and 65 were trapped in another (MORRIS, unpublished), so local densities can be very much higher. Densities in woodlands may also be higher than 1/ha (STORCH, 1978) and it is also likely that Glis has recently spread to new areas. Thus a realistic estimate for the total present British population would be at least 10,000 animals, perhaps more than twice this figure. Since there are about 15,000 hectares of woodland in the Chilterns, this estimate seems a reasonable one and was adopted by HARRIS et al (1995) in their review of the status of British mammals. It is surprising that such a large population has built up in less than 100 years. This species does not appear threatened, on the contrary its population is probably expanding and future establishment in widespread locations is likely. There seems little prospect of its extinction, except perhaps as a consequence of a major climatic change reducing foraging and reproductive success during the already-brief period of summer activity.


Damage to trees seems to have escaped notice until 1963 (PLATT & ROWE, 1964; JACKSON, 1994). Minor damage occurs to species such as Scots pine (Pinus sylvestris) and birch (Betula pendula). Fruit trees and stored fruit are also attacked, but orchards are of declining significance in the Chilterns (JACKSON, 1994) The principal trees affected are Norway spruce 15-30 years old and European larch 28-35 years old. Damage seems mainly to occur in June and July, when the animals gnaw short strips of bark, evidently to eat the cambium layer below. Even minor damage may result in fungal infection of the tree. Often the gnawed patch encircles the tree several metres above the ground and leads to death of the tree crown. In high winds the trees later break off at the point of attack. Four hectares were affected within five weeks (PLATT & ROWE, 1964) and a survey of 14,000 trees in Forestry Commission woods (BRADY, unpublished, 1990; cited by JACKSON, 1994) found 14.6% had been gnawed by Glis, representing a 25% loss in revenue equivalent to 220/ha. In some stands, up to 70% of the trees had been affected and losses were estimate at 2000/ha. Similar damage can be caused by squirrels (Sciurus carolinensuis), whose control is effected by the use of Warfarin, an anticoagulant poison. Although this used to be a means of controlling Glis, it is no longer available as that poison has not been type-approved for use on the fat dormouse.

Levels of damage vary from year to year and this is sometimes interpreted as evidence of fluctuating population density in Glis (JACKSON, 1994). However, the reproductive biology of Glis, so far as it is known, does not lend itself to causing major annual changes in abundance. It is more likely that apparent variations in damage result from different levels of monitoring or perhaps changes in the trees themselves. For example, in some years weather conditions may promote more copious sap flow at certain times. As they get older, thicker bark may also make the trees unpalatable and less liable to attack. The problem seems to be a peculiarly British one; tree damage seems not to be a major issue in the Continental literature (although damage to fruits and nut crops is; SANTINI, 1978). It is not clear why Glis attacks trees in some years but not others. It may be positively attracted by an appealing source of food, or it may resort to gnawing bark when its preferred food is unavailable. This might be the case in years when beech trees do not flower profusely, probably an important food source normally. It is likely that the damage is exacerbated by the fact that plantation trees are all of similar age and if one is affected, then many are likely to be attacked simultaneously as all offer similar potential food at the same time. A similar situation exists in connection with the prevalence of bark damage caused by grey squirrels in British plantation woodland, compared with the natural forests of their native home in North America (KENWARD, 1989). In natural forests, trees of a vulnerable age are often dispersed, hiding the impact of squirrels. The squirrels were found to attack the most vigorous trees in the plantations and many of them. A similar subjective impression was reported by PLATT & ROWE (1964) for the damage caused by Glis. Thus, what is often considered a pest problem' is at least partially man made; ideal conditions being artificially created for extensive damage to a crop species. This seems particularly so where mixtures of beech and conifers have been created, an unnatural condition in Britain, where there is only one native conifer species and this rarely grows in juxtaposition with beech.

Glis also causes significant nuisance in houses, making an unacceptable mess with faeces and urine, a lot of noise at night and by gnawing woodwork and electrical wires. Dormice are alleged to have damaged the organ pipes at Aston Clinton Church.

In the 1995 questionnaire survey (MORRIS, unpublished) a ratio of 2:1 respondents asserted that Glis does cause damage and the damage is done every year not periodically Despite this, opinion was equally divided over whether Glis should be considered a pest or not. This may be partly because 50% of respondents had seen the TV programme "The Incredible Edible Dormouse" which portrayed the animal more as a nuisance than a pest and had also emphasised other qualities such as interest and scarcity. The relatively low return (less than 10%) of questionnaire forms, despite widespread availability and publicity, suggests that the problems caused by the animal do not affect a large number of people. Nevertheless, where they do occur, problems are frequently acute.

Damage to forestry interests is significant, but probably controllable by repeatedly removing the animals from specially supplied nestboxes, to which they are easily attracted. Damage and nuisance in houses may also be serious, but control remains elusive. It is not known why Glis enter houses and further research is needed before the problems can be addressed. Meanwhile householders are constrained by the law, and practical considerations limit what they can do to help themselves. Professional help is available through pest control companies, but is expensive. No research has been done on numbers in houses or whether removal of all (or some) of the animals has any lasting effect. As Glis is likely to have a very sensitive olfactory system, and pheromones could play an important role in social behaviour, there may be scope for development of repellents.


Glis is the only established British mammal (except perhaps some of the bats) for which a full ecological study has not been undertaken and little is known about its basic ecology here. Lack of ecological information makes it difficult to formulate appropriate policies either for its control or its conservation. It is not clear to what extent the animal behaves in Britain as it does on the European mainland, nor therefore the extent to which ecological data from the Continent can be used as the basis for action in Britain. It is unlikely that the animal's biology will be the same in the cool maritime climate of England as it is for example in hot, dry Mediterranean areas 2000km further south. Extrapolations from Continental data are therefore to be made with caution and efforts should be made to obtain comparative data from England.

Trapping has been relatively unsuccessful in the Chiltern forests (PLATT & ROWE, 1964; A. HOODLESS, pers. comm.), so population densities were estimated by HOODLESS & MORRIS (1993) by walking transects through the woodland and counting animals heard calling. Numbers per hectare were derived using the method of WHITESIDES et al (1988). Population densities of between 0.8 and 1.7 per hectare were found. However, this only included calling animals, the actual density may well be higher, especially as PLATT & ROWE (1964) reported 16 occupied nests in a plantation compartment of only l.5ha.

To investigate movements MORRIS & HOODLESS (1992) captured two adult animals, a male and a female, in an outbuilding (formerly a stable for horses, later used as a store for wood and coal). These were fitted with radio transmitter collars (Biotrack, Wareham, Dorset) and their activity monitored for 13 nights. They did not stay in the building, although one or both sometimes returned to pass the day inside. The male ranged up to 250m away, but the female only once exceeded 100m. The mean total distance travelled per night was 523m for the male and lllm for the female. They climbed to heights of 5-8m above the ground to feed mainly in beech and yew (Taxus baccata) trees. The male also used low bramble shrubs (Rubus sp), but the female spent more time in elder (Sambucus niger) bushes. The animals were often seen eating the fruits of these trees and occasionally they also took nuts from hazel (Corylus avellana). They returned to favoured feeding sites for several consecutive nights, then moved to another source of food as the first was used up. It is likely that Glis are mainly arboreal feeders and rely on a sequence of seasonally available foods produced by trees and shrubs, as does Muscardinus (BRIGHT & MORRIS, 1993).

The female went into hibernation surprisingly early, on August 15th. She was found next to a drain outside the building where she was originally caught, and about 30cm underground. The hibernaculum was dug out eight weeks later, during which time she bad not moved. She was in deep torpor and accompanied by a male and another female. The radio-tagged male also went underground (between September 30th and October 4th), within 60cm of the radio-tagged female. He too was dug out on October 13th, and found to be accompanied by a female and another male. Both the hibernacula were cavities in dry soil. There was no nesting material present and no obvious way in. Whatever route the animals had used to reach the place where they were found (perhaps following tree roots or the drain pipes underground), they had filled their access tunnel with soil and were entirely sealed into their hibernacula. In these circumstances it seems almost certain that, in both cases, all three animals had entered hibernation together. A third radio tagged animal (male) also hibernated underground, this time about lm below the surface in a disused burrow probably originally dug by a badger (Meles meles). It was hibernating alone. THOMPSON (1953) reported hibernating animals being found underground, under floor boards in houses and also using a nest of paper on a shelf in an outbuilding.

As Glis are rarely reported active before May, hibernation apparently extends for at least seven months, leaving little time to breed and raise young to adequate size for overwinter survival.

On the Continent, Glis use bird nestboxes readily (GAISLER et al, 1977; SANTINI, 1978), as does Muscardinus in Britain (MORRIS, BRIGHT & WOODS, 1990) enabling collection of basic biological information. For Glis in Czechoslovakia, mean litter size is 4.5 (range 2-8) and survival is at least 3 years in the wild (GAISLER et al, 1977); in Germany, survival to at least 5 years is known (BIEBER, pers comm). For comparison, some trial field studies of Glis have been initiated recently in the Chilterns. Experiments with nestboxes have proved successful, despite their reported failure by PLATT & ROWE (1964). In 1995, between May and July, 90 nestboxes were put up in a wood near Chesham, supplemented in early 1996 by 20 more. They are of several designs, but all are made of untreated softwood, about 1.5cm thick enclosing a cavity some 10-15cm square and about 30cm high. Initially the entrance holes had been 35 mm diameter, but soon many were enlarged to 50-55mm, by gnawing (presumably by Glis, as many of these animals have been found using the boxes, but only 2 squirrels). This seems to be the preferred hole size, as none have been enlarged further.

By October 1995 twenty nestboxes had been used. Often groups of animals were present, often of mixed sexes, although the functional significance of this behaviour is unclear (PILASTRO et al, 1996). Litter sizes in 1995 ranged from 2-5 young (up to 7 young have been reported previously in England; THOMPSON, 1953). The weight distribution of the animals in October 1995 (Fig 2) showed three distinct cohorts. These could represent a parental generation (born in 1994 or earlier), plus early and late litters born in 1995, especially as that summer was long and hot. However, it seems unlikely that young of the year could have grown to 90g or more by September unless they had been born earlier than about June, leaving time for the females to have conceived second litters, comprising the cohort weighing about 20-80g. But early breeding seems unlikely when the animals may be still hibernating in May. In 1996 there were no Glis in the nest boxes in May and numbers did not peak until August. In June, July and August 1996, the majority of animals found in the nest boxes were of this intermediate size (70-150g) and were not evidently breeding (no nestlings were found at all in 1996). The tri-modal size pattern in the 1995 population thus seems more likely to reflect a single litter per year, produced in August-September, with the previous year's young forming the 90-130g cohort, perhaps too small to breed until their third calendar year. However, if this is the pattern of reproduction in England, it seems surprising that such a large British population could have developed in a mere 95 years.

In July 1996 three Glis weighing 71, 120 and 112 g were found together in a nest box in a cold and lethargic state. It is not known whether it is normal for British Glis to enter facultative summer torpor to save energy during inclement weather, but it is common in the native dormouse (Muscardinus).


The basic ecology of Muscardinus has now been investigated thoroughly in Britain, and work continues. However, Glis remains poorly known. The small field studies reported here are the first undertaken in Britain. They provide preliminary information about movements, population density, breeding and hibernacula for Glis living in this country, but many questions remain. For example, why has the animal become so abundant locally but not spread more widely? Why does it attack trees? Why does it enter houses and are some houses more attractive than others? How can we best control Glis in places where it is a nuisance, yet at the same time honour the obligations of the Berne Convention to protect it? Resolving these dilemmas requires a much more detailed understanding of the animal's ecology than we have at present.

Fig. 2. Size frequency histogram showing weights of fat dormice found in nest boxes in October 1995. The three general size cohorts suggest that Glis may not breed before their third calendar year in Britain (see text).


I thank the Royal Forestry Society and the Forestry Commission for permission to work in their woods and for much practical assistance. I am grateful to Paul Bright, Andrew Hoodless, Ruth Temple and many others for help in the field. The Royal Air Force (Halton) provided valuable assistance with accommodation during fieldwork. I am also grateful to Jenny Jones (Hertfordshire Record Centre), lain Bishop (Zoological Museum, Tring), John Jackson (Royal Forestry Society), Emma Lansdell (Buckinghamshire County Council) and the many people who completed questionnaires and supplied further information on Glis in 1995 and 1996. The maps were drawn using DMAP (software Alan Morton).



A review of the fat domouse (Glis glis) in Britain

P. A. Morris

The fat dormouse was introduced to Britain in 1902 by Walter (Lord) Rothschild, who liberated a small number in Tring Park, about 100 km north west of London. The species is still confined to a relatively restricted area, mostly within about 35 km of its original introduction. Here it is often sufficiently numerous to be a serious nuisance in houses and in forestry plantations. Its legal status is confused, being both legally protected, but at the same time an alien species and significant pest. Licences may be granted to trap Glis where they are causing problems, but they may not then be released into the wild because they are not native species in Britain. Many people are unaware of this and increasing numbers of fat dormice are translocated away from problem areas and released (illegally) elsewhere. This is likely to spread the species beyond natural barriers that have hitherto constrained distribution. The total British population is unlikely to be less than 10,000. Under the Bern Convention ' all dormice (Gliridae) are protected and populations are supposed to be monitored, but nobody seems to be regularly monitoring Glis in Britain. Lack of even the most basic ecological data on Glis makes it difficult to act, but some small studies have now begun aimed at understanding more about population density, movements and potential ways of controlling this animal. Preliminary field studies suggest that the animals do not travel very far during their nocturnal activity periods. It is not clear what they eat, but foods are likely to include a seasonal succession of different flowers, fruits and perhaps insects. It appears that Glis may hibernate for up to 7 months of the year (although in some years hibernation may not begin until November). Hibernation is sometimes communal and takes place underground or in a cool sheltered place such as below the floor of a house. Nest boxes are readily used in summer, by individuals, groups of immature animals and by breeding females with their young. Limited evidence so far available suggests that British Glis breed only once a year and may not be mature until their third calendar year. Litter sizes range up to at least 7.