Melanistic Eastern Gray Squirrels - Citizen-Science in My Neighborhood

A year ago, I decided to follow in the footsteps of biologist Richard M. Lehtinen, who, over the course of about eight and a half years, took walks through the neighborhoods around The College of Wooster, in Wooster, Ohio, and tallied each black and gray squirrel he spotted.  The data he gathered were then used as part of a study he co-authored about the distribution of the black (melanistic) color morph of the eastern gray squirrel (Sciurus carolinensis) in the Great Lakes region.  (Dispatches From the Neighborhood Watch:  Using Citizen Science and Field Survey Data to Document Color Morph Frequency in Space and Time, Ecology and Evolution, Volume 10, 2020.)  I discussed this and other research on melanistic eastern gray squirrels in a 2023 post.

Lehtinen walked seven set routes daily in his study area - one route was followed for the entire length of the survey period (July 22, 2010 - December 31, 2018), five others were followed for six years, and one was followed for five years.  He recorded a total of 39,966 observations, 66 percent of which were of melanistic individuals.  These data do not provide an unduplicated count of squirrels sighted, given that many of the same individual squirrels were sighted and tallied on more than a single day.  In the aggregate, though, they do gauge the overall melanistic representation among eastern gray squirrels living in the area and along the routes followed.

It's a fairly simple research activity that generated data important for explaining possibly why distribution patterns for the melanistic squirrels are what they are.  Lehtinen et al. found significant differences in the presence of these squirrels among the routes followed, and that these differences changed over time, some consistently in one direction or the other (increase or decrease in melanistic percentage), and others evincing little change.  These walking survey data, coupled with additional population data assembled by the authors, showed the melanistic distribution in Wooster, across much of Ohio, and in the Great Lakes region to be decidedly mosaic in nature.  In essence, different locations showed sometimes substantially different concentrations of black squirrels, a finding true for areas that were geographically close to each other with similar habitats which, the authors posited, exhibited few, if any, obvious distinctions in selective forces that might be in effect (e.g., presence of predators).  Hence, it would appear that something other than natural selection might be driving the ebb and flow of these black squirrel populations:  the authors suggested that chance in the form of genetic drift was at work.

Beginning in November, 2023, and continuing through December 29, 2024, I conducted my own citizen-science project in which I tallied each sighting I made of a black or gray squirrel during my daily walks through my own extended neighborhood.  My primary goal was quite simple:   estimate, as best I could, the overall representation of melanism in the squirrel population in this area.  A secondary objective was to see how the black squirrels were distributed among several localities within the study area.

The area in question is part of Silver Spring, Maryland, a suburb of Washington, D.C..  The study area is shown roughly outlined below in a screenshot from a portion of the map of Silver Spring from Apple Maps.  My walks fell within this area.

My effort differed from Lehtinen's in several significant ways.  The primary and most important one is that mine covered a much shorter time period than his.  I began tallying my squirrel observations on November 5, 2023, and, for this post, used data up through December 29, 2024, a period encompassing 421 days.  I recorded one or more squirrel observations on 269 days or about 64 percent of the days in this period.  Travel, nasty weather, health issues, and an absence of sightings account for the missing days.  In contrast, Lehtinen's survey was conducted for about eight and a half years - from July 22, 2010 to December 31, 2018 - a period of 3,085 days.  During that time period, he made 5,166 surveys, about 1.7 each available day.

Among other differences in the two neighborhood surveys is the larger area covered by mine.  His routes traversed an area of 0.33 square kilometers or 0.13 square miles.  My routes fell within an area of very roughly 0.73 square kilometers or 0.28 square miles.

In contrast to his effort, I did not have precisely prescribed routes.  As described below, I traversed some streets, avenues, and parks in the neighborhood much more frequently than others.  To minimize double counting of individual squirrels in any one day, the daily circuits I followed through my study area were nearly all loops.  In the event my path did cover any of the ground twice on a particular day, only tallies that exceeded those from the initial passage were included.

Finally, I would stress that the topographical nature of the area being studied in each of these projects may differ.  My study area included several relatively quiet, tree-lined residential roads, and three busy and noisy thoroughfares, among them, one undergoing substantial reconstruction in this period.  Included in my study area were two large parks, one abutting a small woods and one bordered by woods and a creek, and one small, partly treed, park bounded by neighborhood roads.  Lehtinen did not describe his study area beyond noting that "each of [his] routes followed streets in residential neighborhoods adjacent to the College of Wooster campus and Cornerstone Elementary School and are situated along parallel or perpendicular streets."  (p. 1529)  After looking at a satellite view of the presumed area in which he gathered his data, I believe his to be more consistently residential than mine if his routes lay only along an axis running directly from the College of Wooster to Cornerstone Elementary School.  In that case, his study area, unlike mine, featured few, if any, natural or manmade barriers to squirrel movement among the routes.

The additional important issue of identifying color morphs should be addressed.  As I noted in the previous post on these squirrels, analyses by biologist Helen R. McRobie of the genetic basis of the color morphs in the eastern gray squirrel found three distinct such morphs:  gray, black, and brown-black.  The genetic source of the last is a heterozygous version of the allele controlling color.  (Helen Rebecca McRobie, The Genetic and Molecular Basis of Melanism in the Grey Squirrel (Sciurus carolinensis), Anglia Ruskin University, Doctoral thesis, September 2014; and Helen R. McRobie, et al, The Genetic Basis of Melanism in the Gray Squirrel (Sciurus carolinensis), Journal of Heredity,Volume 100, Number 6, 2009.)  To my untrained eye, there is some fluidity in the color variations sported by individual eastern gray squirrels, fluidity that goes beyond the three distinctions McRobie identified.  (See, for example, the discussion of several other color variations "in the vicinity of Washington, D.C.," by zoologists Richard W. Thorington, Jr., and Katie Ferrell in Squirrels:  The Animal Answer Guide, 2006.)  Nevertheless, despite detecting some slight variations in the coloring of a very few of the squirrels I observed, I had no issue distinguishing black from gray individuals in my study area.  All my observations fell into of those two categories.  Lehtinen et al. did not indicate that such color distinctions were an issue for them.

The data provided below are based on tallies of observations on days on which I actually walked the neighborhood and spotted at least one squirrel, whether gray or black.  This means that the few days on which I walked but failed to spot a single squirrel are missing from these data.  The aggregate data provided below regarding the overall representation of gray and black squirrels, as well as the data used to determine differences in melanistic presence in the several specific subsets of neighborhoods I covered (see below), are based on the entire data set of days when squirrels were tallied.

Aggregate Observations

Over the course of the entire study period, I recorded 1,565 observations of squirrels in my neighborhood.  Of this total, 1,152 or 73.6% were of gray squirrels and 413 or 26.4% were sightings of melanistic individuals.

I have no prior baseline data for this or nearby areas with which to compare this melanistic representation in my neighborhood population.  I take with the proverbial grain of salt the unsubstantiated and vague estimate offered up in a recent article in The Washington Post by freelance writer Ashley Stimpson.  She stated that black squirrels make up "as much as half of the District's squirrel population."  (Why We Celebrate Black Squirrels, But Think of Gray Ones as Pests, The Washington Post, October 8, 2024.)  As a source for this estimate, she linked to a CBC News article which posited that "about half of Washington's squirrel population" is melanistic.  This article offered no source for its estimate.  (Matt Kwong, White House, Black Squirrels:  How 8 'Desirous' Canadian Rodents Ended Up Overrunning Washington, CBC News, July 12, 2018.)

Melanism Representation in Several Distinct Localities of the Study Area

Although I did not follow fixed routes during my survey, there are at least five distinct portions of the study area in which I frequently walked.  Some of these localities, though often adjacent to one another, have features potentially posing some physical boundaries (e.g., busy roads) that separate them.  I describe each locality below.

Locality 1:  Two tree-lined, relatively quiet residential roads whose houses share backyards.  The roads are connected by a grassy walkway.  This locality is separated from the others by a busy road, though squirrels do cross it, sometimes unsuccessfully (see data below on squirrels killed by traffic).

Locality 2:  The open and partially wooded area of a park with several ball fields, separated from Locality 1 by a busy street.  Part of the park abuts Localities 4 and 5.

Locality 3:  The broad lawn with mature oak trees along the side of a church, bounded by a very busy road undergoing construction, the parking lot of a small cluster of stores, and a small apartment building.  It seems quite isolated from the rest of the study area.

Locality 4:  A triangular area bounded by tree-lined, relatively quiet residential roads, encompassing a small park with a soccer field and tennis court, partly contiguous to Localities 2 and 5.

Locality 5:  Two, tree-lined, relatively quiet residential roads, that cross each other, partly contiguous to Localities 2 and 4.

My walks took me to each of these localities with varying frequencies.   Of the 269 days on which I recorded any observations, 49.4 percent included observations made in Locality 1, 33.5 percent in Locality 2, 11.5 percent in Locality 3, 26.4% in Locality 4, and 19.7% in Locality 5.  Nearly all of my daily walks ventured into more than one of these localities.   I made observations in each locality in all seasons of the year.

The table below shows the aggregate counts of gray and black squirrel observations in each locality, and the percentage of the total observations in each locality that were of melanistic individuals.

Locality 1 may well have had an outsized influence on the aggregate counts reported earlier.  This can be attributed to the high frequency with which daily observations came from this locality (almost 50 percent), the large number of aggregate observations from this locality (453 or 28.9 percent of aggregate observations across my study), and the healthy representation of melanistic squirrels living in that area (more than 42 percent of Locality 1 observations were melanistic and those accounted for 46.5 percent of all melanistic sightings in my study).

I would note that the data Lehtinen et al. provided in a supplementary data table to their study ("wooster_summary_route_year") show that Lehtinen's Route 1 contributed significantly more data to his survey data than did his other routes.  Approximately 44.0 percent of the total observations from his entire walking survey were accounted for by Route 1 and 46.9 percent of all melanistic observations came from that same route.  No other route he followed came close.  The nearest was Route 2 with 17.0 percent of total observations and 15.7 percent of melanistic observations.  Nevertheless, given the high representation of melanistic squirrels across all  of his routes (the range was from about 49 percent to over 73 percent with a median of approximately 69 percent and a mean of about 66 percent), the aggregate data he reported were not disproportionately influenced by his Route 1.

It would appear that my study showed melanistic squirrels to be distributed across these nearby localities in a mosaic pattern with some substantial differences, often much larger than those from Lehtinen's walking survey data.  Even those localities abutting each other without physical barriers (Localities 2, 4, and 5) showed sizable differences.

It may be that, due to the physical barriers demarcating some of my localities, my findings are somewhat more reflective of the mosaic pattern that Lehtinen et al. found in the separate surveys they conducted of 55 urban areas in Ohio.  The frequencies of melanistic squirrels across these urban areas was "extremely broad."  (p.1534)

I don't know whether the habitats of my localities were similar in the selective forces that might be at work in each.  A few observations are in order.  Several of these localities are bounded by high-traffic roads presenting a threat to squirrels seeking passage across them (particularly Localities 1 and 3).  Over the course of my survey, I counted five dead gray individuals on the road separating Localities 1 and 2.  (I observed one other dead squirrel in this period, a melanistic squirrel inside Locality 1.)  Locality 3 is effectively isolated by the road barriers.  Human traffic and activities, including dog walking, were evident in all of these localities.  Locality 2 also featured well-attended organized sports in the spring and fall.  With regard to natural predators, there are hawks in the area and several were observed or heard in Localities 1, 2, 4, and 5 at various times in the year.

Although more limited in its duration and less structured in its execution than the walking survey by Lehtinen, I believe that the results of my citizen-science effort are not without their utility, offering some support for the findings Lehtinen et al. described in their 2020 study of melanistic gray squirrels in the Great Lakes region.

Eastern Gray Squirrels - Coloring Outside of the Lines

 Earlier this year, I had a question about the coloring of eastern gray squirrels (Sciurus carolinensis).  In this post, I lay out some of what I’ve learned and ask some of the additional questions I have yet to answer. 

In May, I hired a tree service to prune trees around my cottage on the North Fork of Long Island (Suffolk County, New York).  During the work, a couple of black squirrels eluded members of the work crew by racing up the trunk of one tree and leaping onto the branches of another.  I had long ago accepted the presence of those black versions (morphs) of the gray squirrel population living near my cottage, but the tree workers were astounded,  swearing they’d never seen such squirrels before.  (I should get something out the way at the outset, the black squirrels are members of the same species as their gray brethren.)

Is it possible that this tree crew working in and around trees every day on the North Fork of Long Island had never before encountered black squirrels?  Turns out that it is.  The testimony of these workers is one bit of evidence supporting the uniqueness of the population around my cottage.  Unfortunately, I failed to take the time this summer to go and personally search for black squirrels on the North Fork, so I have turned to the citizen-science app iNaturalist.  Observations posted on the app are accompanied by one or more photographs of the specimen in question.  I reviewed over 360 gray squirrel observations from Suffolk County (which includes the North Fork where my cottage is located) that have been posted on iNaturalist.  For all of Suffolk County there are very few black morphs and none that I could find from the North Fork.  (My use of iNaturalist was prompted by a recent study of gray squirrel morphs in the Great Lakes region:  Lehtinen, et al., 2020.  More on that study below.  Full citations to this and other references cited in this post are included in the list of sources at the end of this post.)  I was also pleased to find that there is an iNaturalist project devoted to the black squirrel.  (A project groups observations based on specified criteria.)  Of the project's over 3,000 sightings worldwide, those from Long Island confirm a paucity of such squirrels on the eastern end of the island.  These data show a very health concentration of black morphs on the western end of the island, in and around New York City.  A bit more on this New York City cluster later in this post.

Up until the tree trimming crew reacted as it did, I hadn’t thought twice about the presence of black morphs living near my cottage because I'm used to them, given that they are not uncommon around my home in the Washington, D.C. area.

Pictured below are a few of the black and the gray squirrels that I’ve encountered in the D.C. area.

The eastern gray squirrel matures to breeding age at 10 months and has an average life span of one year.  (Lehtinen, et al., 2020.)  That one year average life span is greatly influenced by a high mortality rate in the initial year; adult females reportedly can live up to more than 12 years in the wild while adult males can live up to 9 years.  (Koprowski, et al., 2016.)   Melanism (presence of dark coloring) is actually quite rare across the entirety of the eastern gray squirrel’s range which encompasses nearly all of the eastern half of the county and parts of Canada (in that range, overall fewer than 1% of these squirrels are black).  At the same time, melanism is common in the northern portion of the squirrel’s range (some 75% of the eastern gray squirrels are black in that stretch of the range).  (McRobie, 2019.)  Of note, the latitudes covered by the northern tier of that range includes Long Island, but not the Washington, D.C. area.  There’s an additional piece of potentially relevant information:  the demes of gray squirrels are small.  (Gustafson and VanDruff, 1990.)  Biologists define demes as local populations of individuals that interbreed, sharing the same gene pool.  This means that these squirrels are likely to interact with only a small group of their squirrel companions.

It's useful, I think, to consider the genetic basis for melanism in S. carolinensis.  Research out of Britain helps ground this story.  Perhaps that origin of this research isn’t all that surprising because the gray squirrel came onto the scene there only relatively recently.  Helen McRobie and her co-authors (2009, p. 709) described the British experience with gray and black squirrels as follows:

Introduced to Britain in captivity in the late 19th century, the gray squirrel has repeatedly escaped into the wild and has subsequently become a successful invader all but outcompeting the native red squirrel (Sciurus vulgaris).  Melanic variants of the gray squirrel are common in North America, but the first sighting reported in Britain was in the early 20th century.

(Later in this post, I describe how I think this description by McRobie, et al., is misleading or, at least, subject to misinterpretation.)

McRobie and her colleagues determined the genetic origins of gray squirrel melanism, finding that there are three color variants among the gray squirrels:  “wild-type gray” (so called to distinguish this typical gray squirrel from the variants), “jet-black,” and “brown-black.”  These color types are related to two specific two alleles (alternate forms) of a gene.  The researchers found that the wild-type grays were homozygous for one of those alleles (that is, in this case, a wild-type gray will have inherited identical versions of this gray allele from each parent and will breed true, producing only gray offspring); the jet-blacks were homozygous for the other allele; and the brown-blacks were heterozygous for these two alleles.  Given the color distributions among these three variants, McRobie et al. concluded that the jet-black allele was “incompletely dominant” to the wild-type gray allele.  (It's a sad commentary on my observation skills that I have lumped the jet-black and brown-blacks together and considered all of them to be black squirrels.)

McRobie (with a different set of co-authors) has concluded that it is most likely that the allele at the core of melanism in gray squirrels did not evolve on its own, but came into this species through interbreeding with fox squirrels (S. niger).  (McRobie, et al., 2019.)

Considering the overall distribution of black morphs in the broader gray squirrel population described earlier (few overall, high percentage in northern latitudes), and the dramatic difference between the population living around my North Fork cottage and those spread across my home territory of Washington, D.C., the question arises:  What accounts for these frequency and distribution patterns?

Research posits that coloring among animals can serve various functions, among them camouflage, signaling, and temperature control.  (McRobie, et al., 2019.)  Various hypotheses about what in the environment might influence the representation of melanism in S. carolinensis populations have been put forward.  The factors suggested and their effects are several and varied.  For instance, some researchers have posited that, in urban environments, black squirrels may be present in high numbers because they are less likely than their gray compatriots to be hit by cars given how conspicuous they are.  (For example, Gibbs, et al., 2019).  Some explain the apparent concentration of melanic squirrels in northern latitudes by arguing it’s a matter of thermoregulation:  dark coloring enables the squirrels to better absorb heat from the sun and, so, survive colder climates.  (For example, McRobie, et al., 2019.  See, also, Thorington and Ferrell, 2006.  Though these latter researchers like the thermoregulation hypothesis, they described the presently available data as mixed in its implications.)  These are quite interesting hypotheses but I don’t see that they explain the differences between my North Fork cottage squirrels and those in the broader Washington, D.C. area.

What if the relative frequency and distribution of melanism in the S. carolinensis population is not a product primarily of any of the various factors suggested to date?  A recent study suggests exactly that.  Biologist Richard M. Lehtinen turned observations he made over several years of eastern gray squirrels on his daily walks in the residential areas around the campus of the College of Wooster (Ohio) into part of the foundation of an analysis of the frequency and distribution of the squirrel’s color morphs in the Great Lakes region.  (Lehtinen, et al., 2020.)  He and his colleagues coupled his observations with data mined from observations posted on iNaturalist for the Great Lakes region, the Ohio Natural History Database (data provided by squirrel hunters), and separate surveys conducted in 59 localities mostly in the northern portions of Ohio.

Melanism was observed across the Great Lakes region, but (and it’s a very important “but”) the distributional and frequency patterns are mosaic in nature.  That is, melanism is highly localized and can vary dramatically from location to nearby location, even though the locations surveyed “appeared highly similar from one town to the next and likely had a similar suite of potential predators and risks associated with human-dominated environments."  (Lehtinen, et al., 2020, p. 1534.)  This means that many of the various factors previously suggested as affecting the distribution and frequency of black squirrels do not explain the mosaic patterns found in this study.  Further, the long-term study of squirrels in Wooster show “inconsistent patterns or idiosyncratic fluctuations over time and space” in morph distribution and frequency.  (Lehtinen et al., 2020, p. 1535.)

How then to account for the mosaic distribution patterns they found?  Lehtinen, et al. suggested that “genetic drift may be an important evolutionary mechanism operating in this system.”  (Lehtinen et al., 2020, p. 1535.)  Genetic drift involves changes in the alleles of the gene that are due to random genetic variation, i.e., chance.  Lehtinen is quoted in an article about his research on the Wooster College website as follows:

These patterns are suggestive of genetic drift as an important mechanism of evolutionary change. . . .  Our initial expectation was that squirrel fur color would matter a lot and that we would see consistent patterns from year to year and place to place.  But because the frequency of black versus gray varies so much from place to place, we ended up concluding that genetic drift – a random mechanism of change – had to be involved.  (College of Wooster, 2020.)

An interesting additional finding was that the frequency of melanism in the squirrel populations studied was highest in mid-latitudes of the Great Lakes region, not in low- or high-latitudes.  So the precise nature of the association between colder climates and melanism remains in question.

As useful as this study is, it has left me with a large clutch of questions, none more pressing than this:  Does the research by Lehtinen, et al. support the idea that melanism can arise quite randomly?  I think it does, suggesting that a local squirrel population of only gray morphs might, over time, come to have black morphs through genetic drift.  But I’m not certain.

The issue of when the black version of S. carolinensis first appears in a population has been muddied by human trading in such squirrels.  I have noted that I think McRobie’s description (quoted in its entirety earlier) of the first appearance of black morphs in Britain is misleading.  She and her colleagues seemed to depict a natural (and random?) advent of melanism in the British gray squirrel population.  That may not be true, given an alternative explanation that points to the escape of North American black squirrels brought over to Britain.  (See, for example, Barkham, 2019.)  Further, is the thinking that all of the black squirrels in Britain have come solely from those North American invaders?  No evolutionary forces at work generating melanism?  I really don’t know, though I’m inclined to doubt it.

Earlier I described data from the iNaturalist black squirrel project which showed a concentration of such squirrels in and around New York City.  I would add that the pattern is decidedly mosaic with some areas in the city sporting many sightings and others not.  Here, too, there's possibly some human meddling.  Michelle Young, in a nice piece on New York City's black squirrels for her website Untapped New York, cited a 1935 New York Times article about a black squirrel being spotted in the New York Botanical Gardens.  In the piece, Raymond L. Ditmas of the New York Zoological Society attributed that black morph to some that had been released several years earlier by the Botanical Garden.  (Young, 2021.)

Perhaps the most well known example of human-moderated movement of black morphs is that of the squirrels in the Washington, D.C. area.  It is widely asserted that black morphs in this area stem specifically from two gifts from Canada of a small number of these squirrels in the early 1900s to the recently established National Zoo.  Released on zoo grounds (of course, they were), these black squirrels soon made their way throughout the Washington, D.C. area.  (Thorington and Ferrell, 2006.)  As a consequence, it seems to be commonly asserted that every black squirrel encountered in this region traces its roots to the National Zoo population.  Indeed, I came across a paper written by a college undergraduate which makes an even broader application of this explanation, suggesting that all of the mid-West’s black squirrels come originally from that population.  (Koleczek, 2014.)  Attributing all black squirrels in the D.C. area to those few individuals shipped from Canada in the early 1900s seems a stretch.  Applying it to all mid-West black morphs seems an impossible stretch.

So, at this stage of this exercise, I’ve concluded that my cottage’s population of black morphs is rare for the North Fork, which explains the tree trimming crew’s astonishment.  How the black squirrels came to be there remains an outstanding question.  And is there some way to implicate the National Zoo black squirrels?

Sources

Patrick Barkham, Black Squirrel ‘Super’ Species?  No, Just a Darker Shade of Grey, The Guardian, August 13, 2019.

The College of Wooster, Rick Lehtinen Examines the Frequency of One of Wooster’s Trademarks – the Black Squirrel – in Recent Study, February, 2020.

James P. Gibbs, et al., The Biological System – Urban Wildlife, Adaptation, and Evolution:  Urbanization as a Driver of Contemporary Evolution in Gray Squirrels (Sciurus carolinensis), in Understanding Urban Ecology, 2019.

Eric J. Gustafson and Larry W. VanDruff, Behavior of Black and Gray Morphs of Sciurus carolinensis in an Urban Environment, The American Midland Naturalist, Volume 123, Number 1, January 1990.

Molly Koleczek, A Survey of the Ratio of Melanistic to Gray Squirrels (Sciurus carolinensis) on 5 Midwestern College Campuses, Honors Scholarship Project, Olivet Nazarene College, April, 2014.

John Koprowski, Karen E. Munroe, and Andrew J. Edelman, Gray Not Grey:  The Ecology of Sciurus carolinensis in Their Native Range in North America, chapter in The Grey Squirrel:  Ecology & Management of an Invasive Species in Europe, 2016.

Richard M Lehtinen, Brian M. Carlson, Alyssa R. Hamm, Alexis G. Riley, Maria M. Mullin, and Weston J. Gray, Dispatches From the Neighborhood Watch:  Using Citizen Science and Field Survey Data to Document Color Morph Frequency in Space and Time, Ecology and Evolution, Volume 10, 2020.

Helen R. McRobie, Alison Thomas, Jo Kelly, The Genetic Basis of Melanism in the Gray Squirrel (Sciurus carolinensis), Journal of Heredity, Volume 100, Number 6, 2009.

Helen R. McRobie, Nancy D. Moncrief, and Nicholas I Mundy, Multiple Origins of Melanism in Two Species of North American Tree Squirrel (Sciurus), BMC  Evolutionary Biology, volume 19, 2019.

Richard W. Thorington, Jr., and Katie Ferrell, Squirrels:  The Animal Answer Guide, 2006.

Michelle Young, The Mysterious Black Squirrels of NYC, Untapped New York, October 14, 2021.