Skip booklet index and go to page content

COSEWIC Assessment and Status Report on the Chimney Swift in Canada


Habitat Requirements


The Chimney Swift spends most of the day foraging for insects on the wing. As a result, it is difficult to associate the species with a single type of habitat; its presence in a particular area largely depends on the availability of suitable nesting sites (DeGraaf and Rappole 1995) and the abundance of insects (Kaufman 1996). Before the arrival of European settlers, the Chimney Swift was associated with old growth forests where their main source of nesting and roosting sites--large hollow trees--were much more common than they are today. Today, the Chimney Swift is found over a large variety of habitats such as cities and towns, villages and rural or wooded areas, but it is most often associated with urban and suburban zones (Chantler 1999; Cink and Collins 2002).

The Chimney Swift is often seen near bodies of water because of the abundance of insects, its main food (Sibley 1988, Sibley and Monroe 1990; Chantler 1999; Cink and Collins 2002). Two studies revealed that three of the five main insect orders consumed were associated with wetlands (Fisher 1958; Fudge 1998). The proximity of nesting and roosting sites inventoried during the Quebec Chimney Swift survey (1998-2002) revealed that 95% of them (140/147) were located less than 1 km from a body of water present on 1:50 000 topographical maps (CWS-QC unpublished data). Among those 140 sites, 90% (127/140) and 54% (76/140) were less than 600 and 300 m from a water source respectively.

Wintering Habitat

The Chimney Swift’s wintering habitat in South America consists of river-edge forest, edge of tropical lowland evergreen forest and second-growth scrub (Rappole et al. 1983; Stotz et al. 1996). It also frequents irrigated farmland and suburban and city centre zones (Hughes 1988). On the Peruvian coast, it regularly occurs to 2500 m and sometimes to 3000 m (Hughes 1988). It roosts in chimneys, crevices, caves (Fjeldså and Krabbe 1990) and hollow trees that are plentiful in the Amazon forest (Whittemore 1981). However, the Chimney Swift’s winter habitat preferences are still not very well known (Stotz et al. 1996; Cink and Collins 2002).

Nesting and Roosting Habitat

For nesting and roosting, the Chimney Swift looks for a dark, sheltered spot with vertical surfaces that it can grip onto and attach its nest (Fischer 1958). Prior to European settlement, it nested and roosted mainly inside hollow trees (living or dead) and occasionally on cave walls and in rocky crevices (Chamberlain 1891; Bent 1940; Tyler 1940; Coffey 1944; Lack 1956; Fisher 1958; Tufts 1986; Godfrey 1986; Erskine 1992). There is a lot of evidence that Chimney Swifts require large diameter trees (> 50 cm diameter at breast height or DBH). In 1985, a balsam poplar (Populus balsamifera) used by Chimney Swifts in the Rimouski area had a DBH of 60 cm (Bélanger 1985) and in 2001, another tree of the same size was also used by swifts in St-Pamphile (CWS-QC Unpublished data). The Vaux’s Swift, a closely related species, also requires large hollow trees for nesting and roosting. Twenty-one trees containing Vaux’s Swift nests in Oregon had a mean DBH of 67.5 cm (Bull and Collins 1993). The birds normally enter the tree through an opening in the top.

The Chimney Swift can also nest in cavities excavated by the Pileated Woodpecker (Dryocopus pileatus), although this is rare (Cameron 1949; Hofslund 1958; Cotrille 1956 in Dexter 1991). Flocks of Chimney Swifts have also been observed roosting on tree trunks. It is possible that birds choose this when there are no other appropriate sites available (Spendelow 1985) or when their usual sites suddenly become unavailable--e.g., when a fire is lit in a chimney (Campbell and Campbell 1944), or because of sudden poor weather conditions that force them to seek shelter (Arvin 1982).

With the arrival of Europeans in North America, forest clearing began and large trees became increasingly scarce (Leverett 1996; Drushka 2000). At the same time, artificial structures (chimneys, barns, wells) were being built, and Chimney Swifts rapidly adopted suitable ones for nesting and roosting (MacNamara 1918; Coffey 1936; Lack 1956; Fisher 1958; Johnsgard 1979; Bull 1985; Norse and Kibbe 1985; Sibley 1990; Peterjohn and Rice 1991; Sutcliffe 1994; Fleckenstein 1996; Snow and Perrins 1998; Cink and Collins 2002). Among these structures, chimneys are the most abundant and by far the most frequently used. Chimney Swifts appear to have adopted chimneys quite early on since they were first spotted in such structures in Maine in 1672 (Palmer 1949). Coffey (1944) mentioned that swifts began using chimneys in 1808. At the beginning of the 19th century, Audubon (1840) had already observed the widespread use of chimneys for nesting. He even commented that the species once nested in trees in western Kentucky, implying that the use of natural sites was already a phenomenon of the past by that time. In the same period, Wilson (1812) observed that nesting was already limited solely to chimneys in western Pennsylvania.

Chimney Swifts choose unused chimneys to roost or build their nests, but a moderate amount of heat does not appear to harm them in large chimneys (J. Gauthier, pers. obs.). Little is known about the factors that contribute to the swifts’ decision to choose one chimney over another but temperature seems to play a role. During the Quebec Chimney Swift Survey, the temperature inside a few chimneys occupied by swifts was measured. The data show that the inside temperature fluctuated very little in relation to the outdoor temperature (CWS-QC, unpublished data). Tyler (1940) reported that the chimneys most frequently occupied were unused, connected to the basements of their buildings and provided a flow of warm air. Bowman (1952) gives an example of such a chimney in Kingston, Ontario, adding that the flow of warm air made the chimney particularly attractive to swifts, particularly on cool April and May nights. In Lévis, QC, in spring 1998, during a cold day, swifts attracted by recorded bird calls did choose a chimney connected to a house over the artificial chimney made out of wood, which did not retain heat (Garneau and Gauthier, CWS-QC Unpublished data). This experiment was repeated many times with similar results. In Quebec, swifts look for sites where the ambient temperature will remain relatively constant and where some heat is present. Garneau and Gauthier (CWS-QC Unpublished data) were able to determine that the threshold temperature at which swifts abandon a chimney is 13 °C. It is likely that large hollow trees also offer this minimal temperature.

In addition to chimneys, Chimney Swifts can nest and roost in air shafts, silos, wells, inside barns, tobacco curing sheds, abandoned buildings and large concrete sewer pipes (Fischer 1958; Bull 1985; Dexter 1991, M. Robert, pers. comm.). Inside buildings, the birds generally build their nests above the floor in the darkest corners (Fischer 1958) where heat tends to accumulate.

Today, most authors suggest that Chimney Swifts nest in chimneys and similar structures, and are less frequently observed using very scarce natural sites (MacNamara 1918; Coffey 1936; Lack 1956; Fisher 1958; Johnsgard 1979; Bull 1985; Norse and Kibbe 1985; Sibley 1990; Peterjohn and Rice 1991; Sutcliffe 1994; Fleckenstein 1996; Snow and Perrins 1998; Cink and Collins 2002). In Quebec, the Chimney Swift roosting and nesting inventory (1998-2004) revealed that only 4 out of the 222 sites were not chimneys, although this ratio may be biased by searches directed towards chimneys and the difficulty in locating natural sites.

As a result, the Chimney Swift is now highly dependent upon humans for nesting sites. In New York State, Sibley (1990) notes that almost all of the swift records during work on the Atlas of Breeding Birds in New York State were in cities or towns. In Ohio, Beissinger and Osborne (1982) observe that the Chimney Swift population density is five times higher in cities and towns than in forested areas. In Rhode Island and Tennessee, the highest densities of swifts are found in urban areas (Enser 1992; Nicholson 1997). However, some swifts probably continue to nest in hollow trees in isolated wooded areas (Fischer 1958; Helleiner 1987; Sutcliffe 1994), but reports of such behaviour are now rare (Norse and Kibbe 1985; Chantler 1999) and the number of Chimney Swifts breeding in forested areas probably represents a very small fraction of the population. Data from the Ontario Breeding Bird Atlas point counts (see Figure 2) show small populations of Chimney Swifts associated with areas of older forest (A. Dextrase, pers. comm.) and Erskine (1992) shows a distribution across New Brunswick, including areas some distance from urbanized centres. Erskine (1992) also states that, in the Maritime Provinces, “…most swifts originally nested, and still nest, in hollow trees.”

It is important to note that nesting sites are difficult to locate because of the very secretive behaviour of the swifts as they approach the nest. Roosts are easier to identify because of the larger number of birds involved, but no roosts in hollow trees have been reported to us, probably because of the scarcity of large-diameter trees. The few birds seen in Saskatchewan have been in remote areas and are probably using hollow trees for nesting and roosting (A.R. Smith, unpub. data). In 1958, Fischer said that the number of reports of Chimney Swifts nesting in hollow trees had fallen considerably since the 1920s. Blodgett and Zammuto (1979) noted that barely 10 nests in hollow trees had been reported in the previous hundred years. Our literature review found only 22 reports of hollow tree nests in the United States between 1840 and 1991 (Audubon 1840; Ridgeway 1874; Daniel 1902; Stewart 1975; Blodgett and Zammuto 1979; Hall 1983; Bull 1985; Robbins 1991; Ferguson and Ferguson 1991; Nicholson 1997). In Canada, Peck and James (1983) report only one such nest in Ontario. In the Maritimes, there are no records of nesting in a hollow tree for New Brunswick, but there are 10 for Nova Scotia (A.J. Erskine, pers. comm.). However, most of these sightings are from one nest, observed during several consecutive years. In Quebec there are six hollow tree nesting cases known (Quebec Nest Record Card Program, Desgranges 1964, Bélanger 1985, Quebec Chimney Swift Survey). Swifts were also observed flying over old forest habitats in Quebec in 2000 (F. Morneau, pers. comm.), 2002 and in 2004 (CWS-QC Unpublished data). The observations made in 2002 and 2004 were part of old growth Chimney Swift survey, where six forests were visited.

Like the Chimney Swift, the closely related Vaux’s Swift also nests and roosts in chimneys. However, this species is still found more frequently using large hollow trees (Bull and Collins 1993). A positive association was found between the Vaux’s Swift and old growth forests (Manuwal and Huff 1987). Pough (1957) stated that the Vaux’s Swift had just started to make the transition from hollow tree nesting to chimney nesting that the Chimney Swift made many years ago.


Many authors have suggested that Chimney Swift populations may have increased with the arrival of European settlers and the multitude of nesting cavities provided by chimneys (Tyler 1940; Norse and Kibbe 1985; Dexter 1991; Kaufman 1996; Zucker 1996; Chantler and Driessens 2000; Cink and Collins 2002). Chimneys and other manmade structures were supposedly more abundant and available than hollow trees present before settlement, thus increasing the number of nesting and roosting sites. These new nesting sites were rapidly adopted by Chimney Swifts.

Graber and Graber (1963) are often cited to support the hypothesis that Chimney Swifts benefited from colonization. They noted an increase in Chimney Swift density for Illinois between 1906-09 and 1956-59 and attributed these results to an increasing human population and development. However, these results do not reflect the situation before and during colonization, but instead represent an urbanization process--10 of the 14 million acres of Illinois’ forest had been cut during the 19th century and by 1900, 33 of Illinois’ 36 million acres had already been modified (Graber and Graber 1963).

It is perhaps just as plausible to suggest that European colonization reduced the Chimney Swift population in North America, since surveys of remnant old growth forests suggest that the number of hollow trees removed was almost surely greater than the number of chimneys built after this event. McGee et al. (1999) found an average of 18 snags (at least 50 cm DBH) per hectare in old growth deciduous forest of New York state. Similarly, Goodburn and Lorimer (1998) found similar results for deciduous old growth forests in Wisconsin and Michigan State (20 snags/ha with a least 45 cm DBH). A rough estimate of 0.152 chimneys/ha in the eastern United States in 1900 can be calculated using United States Census Bureau (2004) data and assuming one house per four persons and two chimneys per house. This figure, though rough, is two orders of magnitude less than similar estimates of snag density before colonization.

In brief, chimneys were not constructed at the same rate as large hollow trees were felled. In Canada (Maritimes, Ontario and Quebec), it appears that the situation was quite similar. The number of households, and therefore chimneys, was less than in the United States, but logging activities and land clearing were of the same order (Historical Atlas of Canada 1990). In the Maritimes, few forests escaped the human influence after the arrival of the Europeans (Loo and Ives 2003). In southern Ontario, almost all old growth was eliminated for agriculture and logging purposes (Suffling et al. 2003). The situation was much the same in Quebec where logging activities went well beyond the inhabited areas of the St. Lawrence River (Dupont 1995).

The number and suitability of remaining chimneys continues to decline. The growing use of electric heating starting in the 1950s was the beginning of the end of this artificial habitat, a process continuing with conversion to heating by natural gas. Today, new buildings have either no chimneys or have metal flues unsuitable for Chimney Swifts. These flues are also often quite narrow (< 30 cm), which can turn them into fatal traps for birds that venture inside because they are unable to get out again. Insurance companies also encourage owners to have metal liners installed in their brick or stone chimneys during renovations as a fire prevention measure. In addition, if a chimney is no longer used to heat a building, the top is frequently capped or the chimney is demolished. In a number of municipalities, fire prevention bylaws oblige residents to install spark arresters in their chimneys. This widespread practice effectively blocks birds and other animals from using chimneys.

Apart from the efforts made in Quebec, there are no quantitative data on the proportion of capped chimneys. However, it is sufficient to observe building roofs in the province to realize that a very large proportion of chimneys are metal or contain spark arresters. Savard (2000) mentions that, in some neighbourhoods in Chicoutimi where the Chimney Swift was recorded in the past, chimneys have been systematically capped or converted and the species seems to have deserted the city.

Suitable chimneys have a large enough diameter (> 28.5 cm), a rough inner surface (e.g., brick, cement, tile) and must offer protection against low temperatures. Chimneys that meet these criteria were generally built before 1960. After 1960, electricity became the prime energy source. Later, during the 1980s came the high performance combustion stoves. These stoves increased the amount of creosote accumulation, which when in contact with water, creates powerful acids, which in turn cause the chimney to crumble. To solve this problem, metal chimneys and metal lining inside traditional chimneys were installed. Terra Cota, the most important clay lining company in eastern North America, shut down in 2001 after more than 100 years in business. The company president and engineer, M. Gaillardetz said that this was a direct result of the change in technology. New technologies also brought smaller chimneys, unsuitable for swifts. Based on sales records of the Terra Cota Company during the 1990s, the small clay tiles used for the smaller chimneys, increased from 20% to 80%. At this rate, it is easy to conceive that most chimneys will no longer be suitable for swifts in 10 years at most, except for the few strong and more resistant chimneys on churches and religious buildings.

Abandoned chimneys often do not offer the appropriate protection for swifts against weather. Once a chimney reaches a certain point in deterioration, it no longer protects against the wind, thus lowering the internal temperature and causing the birds to leave the site if it falls beyond a certain point. Such chimneys also represent a safety hazard and are usually quickly destroyed, especially in the case of industrial, commercial, and government buildings.

Chimneys from residential buildings are for the most part already unsuitable for swifts. About 75% of residential chimneys either possess a metal tube inside or some sort for a cap at the top (fence, spark-arrester or hat) (personal communications from P. Allard of the Giroux-Maçonenx brique et pavé Company, M. Gaillardetz of the Terra Cota Company, Fire Department of Montreal and the Professional Wood Heating Association). This high percentage of unavailable chimneys is easily confirmed by simple observations in cities and rural areas; of the 25% left, almost 60% have a diameter of 28.5 cm or less (M. Labrecque chimney sweeper and builder, pers. comm.), making them less preferred by swifts, which have a mean wingspan of 30 cm. In such cases, swifts need to crawl out of the chimney.

Last but most important are chimneys from churches, rectories and neighbouring schools built before 1960. Most are impressive structures made out of either bricks or stones, and are more resistant than residential chimneys. According to Dr. R. Pleau of the Architecture Department at Laval University, such chimneys with the newer cement have an average lifespan of 60 years. Although many are not used by the Chimney Swifts, they seem to be the most preferred sites. This may simply reflect the reality that these chimneys have not yet been rebuilt; the high cost of such renovation, the high number of buildings and the decreasing popularity of religion make these chimneys a low priority for building owners. Religious-building chimneys represent 57% of all known sites (nests and roosts) and 79% of all known roosts in Quebec (Gauthier et al. in press).

Since 1998, the Quebec Chimney Swift Survey has found swifts inside 40% of church or rectory chimneys (Gauthier et al. CWS-QC, unpublished data). In light of this, we conducted a study to estimate the proportion of church and rectory chimneys that are still available in Quebec parishes, looking at parishes founded prior to 1960. The results of this study appear in Table 1. Approximately 35.4% of chimneys (131/370) in the parishes selected are no longer available for swifts. The diocese with the highest rate of closure is Montréal, where 54% of church and rectory chimneys have been closed. The most common reason for the chimneys’ unavailability was the installation of a spark arrester, a hat or protective fencing (Table 2). There is every reason to believe that the situation is similar in the rest of Canada and in the United States, although the rate of closure and conversion is probably different due to different weather conditions.

Table 1: Proportion of Church and Presbytery Chimneys Unavailable for the Chimney Swift in Quebec Dioceses
DiocesesNumber of Parishes SelectedNumber of Chimneys SampledNumber of Chimneys Unavailable(%)
St Jerome, Joliette20321753.1%
Quebec City33432251.1%
Chicoutimi, Baie Comeau22401537.5%
Nicolet, Trois Rivières24391435.9%
Valleyfield, St-Jean Longueuil, St-Hyacinthe37531426.4%
Amos, Rouyn-Noranda1318422.2%
Gaspé, Rimouski and Ste Anne de la Pocatière3654814.8%
Mont Laurier, Gatineau–Hull1521314.3%


Table 2: Changes to Church and Presbytery Chimneys in Quebec that have Rendered them Unusable by Chimney Swifts
ModificationsNumber of Chimneys(%)
Presence of a spark arrester, hat or protective fencing6650.4%
Metallic flue inside the chimney2317.5%
Prefabricated chimney1713%
Chimney capped1612.2%
Chimney demolished96.9%

In Quebec, there are 1605 parishes that were founded prior to 1960 (Anonymous 2000). If we estimate that there are three chimneys per parish, one for the rectory, one for the church and one for the elementary school, which is very often located next to the church in old parishes, we obtain a total of 4815 potential chimneys for the Chimney Swift. However, this number is probably lower because some churches do not have a chimney or the rectory is in the same building as the church. Some parish churches have more than two chimneys, but if the percentage of closure obtained with our sample (35.4%) is applied to all of the chimneys, we arrive at a total of 1704 closed chimneys. This would leave 3111 church and rectory chimneys potentially available in Quebec parishes, from which many may not be suitable or even available to swifts. If the last chimneys were constructed in 1960 and the maximum lifespan of these chimneys made out of bricks and cement is 60 years, then by 2030 very few traditional chimneys will be left, and many of them will disappear in the next 5 to 10 years. At that point, the Chimney Swift will face a severe shortage of nesting and roosting sites. Although chimneys on religious buildings are not the only breeding sites available for this bird, they probably represent the majority.

According to Simard (1998), Quebec’s religious heritage (architecture, landscape, furnishings and archives) is threatened and has become increasingly impoverished and degraded over the past few years. Accordingly, the Quebec government has invested $101.5 million since 1995 for the restoration of this heritage, particularly church renovations (Government of Quebec 2000). Under the program, 18 churches will undergo major restoration work in 2000 (Government of Quebec 2000). In half of these cases, the roofs will be repaired and it is highly probable that the chimneys will be renovated at the same time. These renovations are not likely to benefit the Chimney Swift. As a result, the rate of closure of church and rectory chimneys could be quicker than expected.

The situation is thought to be similar in the rest of Canada and in the United States. The rate of chimney conversion, destruction and closure is probably faster in the more northern latitudes because of climate, which could explain why Chimney Swifts are decreasing faster in Canada than in United States.

To conclude, using data from Statistics Canada (Building and Labour census 1951), it is possible to estimate how many potentially suitable chimneys are left. Most of the suitable chimneys left are on churches and religious buildings and residential homes. Since the majority of chimneys in the 1950s were not suitable for swifts, as a result of being used by cooking stoves, we must refer to more affluent households which had two chimneys. To this total we must also add households that converted to electricity with their chimney remaining potentially available for a certain period of time. Today, about 75% of these chimneys are not available to swifts. With a detailed analysis of Statistics Canada’s figures, it would be possible to determine the potential of available chimneys for swifts today. Based on our first estimates, this potential is low.


The environment in which the Chimney Swift lives makes it difficult to link it with the concept of habitat protection as usually defined. A large proportion of nesting sites used by Chimney Swift are not protected, because they are chimneys on private buildings. There are less than 10 known roosting sites in the Maritimes and two of these--one in Fredericton, NB and another in Wolfville, NS are well known and are under the protective care of local volunteers. In Ontario, there are no specific arrangements in place for the actual conservation of chimney roost sites, but some of the building owners are at least aware of the roost on their property. In Quebec only nine urban sites are protected in some way and could be eligible for Environment Canada’s stewardship program. They are six roosts located in Chantler, La Pocatière, Mont-Laurier, St-Raymond de Portneuf, St-Georges de Beauce and Joliette, as well as three nesting sites in Joliette, Lévis and Mont Mégantic. The owners of these nine sites are aware of the birds’ presence and efforts are made to maintain the chimneys’ availability.

There are probably few nesting sites in forests, since snags, hollow and sick trees are usually eliminated during harvest. In the Maritimes, only 1 to 5% of the forest cover is presently old growth (Mosseler et al. 2003). However, Nova Scotia aims at conserving 8% of its Crown forest land toward achieving and maintaining old growth conditions (Nova Scotia Dept. of Natural Resources 2004). For New Brunswick, this figure reaches 19% of crown land (D. Beaudette, NB Dept. of Natural Resources, pers. comm.).

In southern Ontario, as of 1986, only 0.07% of the land south and east of the Canadian Shield was classified as greater than 120 years old (Larson et al. 1999). In central and northern Ontario, the percentage of the forest in old growth condition is considerably higher, averaging 23% in Crown forests, and 28% in parks and protected areas (OMNR, 2002). As in Quebec, hollow, dead or dying trees are often removed during logging for safety reasons. However, current silvicultural guidelines within the range of the Chimney Swift in Ontario (Naylor et al. 1996, OMNR 2000, 2001) include the maintenance of (usually 6) large, live cavity trees or potential cavity trees in every hectare of managed forest. However, cavity trees are defined as having a healthy crown (OMNR 2000) and may not be suitable for Chimney Swift. Large, dead trees used by roosting Pileated Woodpeckers, though rare, are to be retained in managed forest on Crown land (Naylor 1996).

In Quebec, seven of the 49 identified old growth forests are classified ecological reserves and are thus protected: the Rivière du Moulin, Tantaré, Lake Malakisis, Tapani, Rolland Germain, Grands Ormes and Boisé des Muir ecological reserves. The old growth forests in these reserves represent 1395 ha or 20.9% of the total area of the old growth forests identified (Government of Quebec 1996).