Skip booklet index and go to page content

COSEWIC Assessment and Status Report on the Eastern Flowering Dogwood in Canada

Population Sizes and Trends

Search Effort

Initially, all known information was brought together, including the records at Ontario’s Natural Heritage Information Centre, field notes from botanists working in the area (Mike Oldham, Jane Bowles, Allen Woodliffe, Gerry Waldron), current herbarium records where available and earlier records the author and associates compiled at the University of Guelph Arboretum. Norfolk County has several recent reports that record the presence of this species (Gartshore et al., 1987; Draper, 2002; Ambrose & Waldron, 2004). Knowledgeable individuals were also queried for specific regions; sometimes they also participated in the field survey.

Populations were sampled across the range of distribution in southern Ontario. An effort was made to visit all populations at the extremities of distribution and sample representative populations in the mid-regions. Where populations were small or in a limited area, an attempt was made to document each individual. In larger areas, such as Skunks’ Misery, St. Williams and Backus Woods, only enough of the site was sampled to get a sense of the impact of dogwood anthracnose on the population.

In the spring and summer of 2004, 17.5 person-days (approximately 140 person-hours) were spent inventorying populations, collecting samples and assessing dogwood anthracnose (DA); two days in the field were spent with Dr. Richard Wilson, provincial forest pathologist. Symptoms for DA were not visible until mid-summer. Samples of 25 populations were sent to Chuck Davis (Fungal Taxonomy and Ecology Technologist) at the Great Lakes Forestry Centre, Sault Ste. Marie, for determination of DA presence. In early May 2005, the two EMAN (Environment Canada’s Ecological Monitoring and Assessment Network) plots in Norfolk County were re-inventoried, with the help of about a dozen students taking a week-long monitoring workshop (8 person-days, approximately 64 person-hours). This early summer 2005 data set was used to give one more year of observational data for the charts on Figures 4 and 5. One more day was spent in the summer of 2005 to check additional sites. In total, 18.5 person days were spent in the field, in 2004 and 2005, not counting the class inventorying the two EMAN plots.


Recent records (1975-2005) indicate the presence of 154 populations in southern Ontario within 12 counties and regions.

Two historical records for Brant County (Fox & Soper, 1952) had imprecise geographic coordinates and could not be found; nor were they known by a local authority who also searched for this species (Larry Lamb, pers. com.).

Of a total of 180 recorded sites (Table 1), 32 were visited and surveyed; 468 living individual trees were recorded at these sites. Another 15 sites with precise counts were recorded in the last five years by Allen Woodliffe and Jane Bowles, for an additional 134 trees. Ten sites were recorded in Norfolk County in 2003 (Ambrose & Waldron, 2004) but not with precise counts; estimating an average from memory of 6 trees per site gives an additional 60 trees. A few sites were visited where the trees were known in the past and they are now absent, due to development or loss. Several more were records with imprecise locations and were not found. In all, of the 180 sites, 57 were observed in the past five years with a total of about 662 trees.

Of the 88 records from 1975-1994 plus an additional 9 recorded from 1995-2001 but not included in the previous total, it is likely that many of these 97 sites are still extant, but how many one can only roughly speculate on. Of the 50 sites cited in Soper & Fox (1952), 76% are still extant (Table 1). If we use 75 as the percent of sites being extant of the above 97, then 73 likely remain. If we assume the same average number of trees per site (12, based on the 662 trees at 57 sites), then another 873 (12 X 73) trees likely exist in Ontario, for an estimated total of 1535 for all 154 sites known between 1975 to present.


Table 1: Summary of Ontario Recorded Sites of Cornus florida
County/regionup to 1952*1953-19741975-19941995-presentTotal
Brant2 (2)0002
Elgin5 (2)16110
Hamilton4 (3)0014
Kent [Chatham-K]327918
Niagara16 (2)5361356
Norfolk9 (3)451830

* Fox and Soper, 1952; number listed (numbers of sites not seen since); each subsequent time period lists new records only (but includes the extant Fox & Soper records). This is a summary of the confidential appendix.

All of the small populations and most of the large populations in small areas (e.g., Cedar Creek, Decew Falls, Fireman’s Park) were sampled thoroughly, but others where trees were scattered over large areas (e.g, St. Williams, Queenston) were sampled only in part. For this latter group, there very likely are additional uncounted trees, perhaps adding another 50 or 100, so that the total living Ontario population is estimated to have been in the realm of 1600 +/- 50 a few years ago.

With an annual mortality rate of 7-8% per year (Wilson Tract and Backus Tract combined mortality of 86.5% over 11 years of monitoring), these numbers could have been reduced by about 20% over three years, with perhaps a more realistic figure of about 1260-1300, and declining (considering that some of the records are 2-4 years old and the recent inventory included trees that were dying).

With reduced flowering and fruiting (Table 2) and only rare sightings of seedlings, recruitment appears to be in decline as well.


Table 2: Summary of Extant Populations of Cornus florida in Ontario, inventoried 2004-05
Co/Reg & site nameMatureLive/
Individuals with flowers/fruitDogwood anthracnoseObserved mortality
Elgin: Springwater11/0+(–)0%
Essex: LaSalle11/0+0%
De Marco21/150%(–)50%
Canard River KCT Woods11/0100%(–)0%
Cedar Creek-Arner169140/2946%+17%
Cinnamon Fern Woods3526/932%(+)29%
Leamington Wh.Oak Wds63/30%(+)50%
Kopegaron Woods4631/1517%(+)33%
Halton: Sassafras Wds106/430%(+)40%
Bronte Creek1110/182%(+)9%
Hamilton: Princess Point11/0+(–)0%
Chatham-Kent: Wheatley Prov. Park2722/519%(+)19%
Wheatley Prov. Park East3224/828%(+)25%
Middlesex: Ausable Valley99/044%0%
Skunks’ Misery1110/127%+9
Niagara: Navy Island43/10%+25
Queenston, E of quarry11/0++0%
Firemen’s Park & N7552/2345%(+)30%
3km E of Thorold43/125%(+)25%
M.A. Sams11/0+(–)0%
Decew Falls to Short Hills3732/548%+14%
N of Fonthill33/0100%(–)0%
Nettle Creek44/075%(–)0%
Balls Falls55/0100%+0%
Norfolk: Wynia64/217%+33%
De Cloet Woods2417/721%+29%
Wilson Tract EMAN plot*84/40%(+)50%
Backus EMAN plot*1713/420%(+)24%
Backus, streambank33/033%(–?)0%
St. Williams, SW2925/459%(+?)14%
St. Williams1410/464%(+)29%
Pine Grove22/0100%(–)0%
mean: 46%
(populations with more than 1 tree)
8 confirmed, 12 likely0-50%;
mean: 17% average: 17.9%

* for EMAN plots, figures between 2003-2005 used, to be comparable with one-time observations for other sites in 2004-5
+Inventoried by J.D. Ambrose and others
Dogwood anthracnose: + = confirmed, (+) = conspicuous symptoms present but not confirmed,
– = no symptoms, (–) = no symptoms but collected early in summer when symptoms are not typically present

Fluctuations and Trends

Since the arrival of dogwood anthracnose in Ontario, perhaps in the early 1990s, Cornus florida has been observed in decline, both through casual observation (Mary Gartshore, pers. comm.) and through monitoring of specific stands (Brian Craig, in NRC, 2004; pers. comm. and diagrams produced for Figures 4 & 5 below). The value of the EMAN plots is that they document the mortality of tagged and mapped trees over an eleven-year period. For the sites visited for this report, dead flowering dogwood trees were also recorded, but they can be determined with some certainty only up to about 2-3 years after death. Trees that died earlier may no longer be standing, or if they are, the diagnostic dogwood branching pattern and bark is likely missing. The percentage of identifiable dead trees in inventoried populations (Table 2) ranged from 0 to 50% (average about 18%; mean 17%); this agrees with the earlier calculated (based on the two EMAN plots) average of 7-8% decline per population per year.

Decline for the whole Ontario population is variable but dogwood anthracnose and tree mortality appear to occur in most areas. Its presence has been confirmed in the following counties/regions from submitted samples: Essex, Halton, Chatham-Kent, Middlesex, Niagara and Norfolk (Chuck Davis and Richard Wilson, pers. comm.). A few isolated areas so far seem to be free of dogwood anthracnose symptoms (e.g., Springwater, Elgin Co. and the Ausable Valley), but other isolated sites are infected, e.g., Bronte Creek and Navy Island.

Figure 4: Comparison of Proportion of Live and Dead Cornus florida in Backus Woods EMAN Plot 1995-2005

Figure 4: Comparison of proportion of live and dead Cornus floridain Backus Woods EMAN Plot 1995-2005.

Figure 5: Comparison of Proportion of Live and Dead Cornus florida in Wilson Tract EMAN Plot 1995-2005

Figure 5: Comparison of proportion of live and dead Cornus floridain Wilson Tract EMAN Plot 1995-2005

Diagrams by Brian Craig, based on his EMAN monitoring study.

For the two sites monitored between 1995 and 2005, mortality counts are cumulative, i.e., in Figure 4, the 38 dead trees for 2003 include the 9 counted in 1995, and the 42 in 2005 include the previous 38 dead trees. Because each tree was tagged and mapped, its condition was tracked during each inventory and compared with its previous condition. The percentage of mortality increased from 14 to 91% for Backus and 19 to 84% for the Wilson Tract. Together, an average 86.5% decline over 11 years was documented (when combining the number of trees vs. averaging the two mortality rates). While this at first seems more drastic than other populations, if dead trees can only be identified to species for 2-3 years after death, then the mean value of 17% dead trees of all inventoried sites agrees with the individual year inventory values for the EMAN plots of new deaths plus living trees.

The continuing loss of trees between 2003 and 2005 at the two EMAN plots does not show an optimistic levelling-off of mortality. Furthermore, among the few survivors in these two plots almost no flowering was observed. 

Only one population can be confirmed to be lost, due to residential development in the Windsor area. Other historical records with imprecise locations could not be relocated.

Rescue Effect

Because dogwood anthracnose-the single factor that seems to be the cause of significant decline in this species-is thought to have originated in the United States, bringing stock from there is not a viable option. Bordering populations as well as those further south are heavily impacted by this disease (NRC, 2004; Schwegman et al., 1998; Anderson et al., 1994). In addition, stock from the south would be less well adapted to conditions in Ontario.

Migratory forest birds likely carry seeds of Cornus florida on occasion, but in the autumn they are more likely going from Ontario to south than vice versa, so only in a very rare event, as in periods of extreme fluctuations in temperature, is recruitment from long distance dispersal likely. However, even though the Carolinian forest habitat is highly fragmented, there is still considerable habitat in which seeds arriving from a long- distance dispersal event could become established.

One could hope to find dogwood anthracnose resistant individuals as part of a strategy to assist recovery, but so far there is little apparent evidence of resistant individuals in stands of infected trees observed in Ontario.