Thanks to my undergraduate mentor, Professor John Moss at Franklin and Marshall College, who got me started in my love of everything glacial, I began my graduate studies at the University of Michigan under Professor Don Eschman. Michigan at that time had a wonderful, diverse program in Pleistocene studies. In my first year I took a course in palynology with Botany Professor Bill Benninghoff. Not only did Bill put me in contact with the Icefield Ranges Research Project in the Yukon where I spent two summers on the Kaskawulsh and Hubbard glaciers for my Master’s thesis dissertation, he gave me an understanding of vegetation associated with the receding Pleistocene ice margins. One of Don’s colleagues was Professor John (Jack) Dorr, a vertebrate paleontologist whose entire career was involved with a search for fossil evidence for early human migration into North America via an interior migration route between the Cordilleran ice sheet and the Laurentide ice sheet. Under the tutelage of Don, Bill, and Jack, plus Profs Bill Kelly and Ed Goddard, my PhD dissertation became a study of the late Pleistocene of southwestern Alberta. My focus was on the evidence related to the deglacial history of the Cordilleran and Laurentide ice sheets, which clearly established that these two ice sheets had coalesced and remained joined until late glacial times, dispelling the theory that early humans initially had migrated into North America via an interior open, ice-free corridor, contrary to Dorr ‘s belief. Nevertheless, as personally disappointing as my findings were to Jack, he signed off on my Dissertation.
Early human history thus has long been a personal interest. In general, the prevailing theory was that late Pleistocene human migration was driven by the availability of large game, such as musk ox, caribou, wooly mammoth and mastodon, that such animals could be taken relatively easily and thus served as a convenient and important food source, and further that such animals shadowed the shifting grasslands and woodlands associated with both the advancing and receding ice sheet margins. Thus, early human migration may as well have shadowed the receding ice margins. Early humans were effective killing machines, decimating game supplies, and thus were continuously migrating in search of food for the large game species which were relatively easily taken by spears, using stone projectile points.
By classical thinking, the earliest was a Paleoindian Clovis culture, characterized by fluted projectile spear points, generally considered to be about 12,500 –13,500 years BP. The lithic derivation of Clovis projectile points provides evidence for alternative possible migration routes.
Mammoth and mastodon remains have been found in Pleistocene deposits in various places in eastern North America, including in Vermont, Quebec, New York, and Nova Scotia. The following table gives some example C 14 dates, thanks to ChatGPT:
The following ChatGPT provided map shows mammoth finds in North America:
And for mastodons, again from ChatGPT:
Obviously, these species were widespread. In general, mammoth diets were predominantly grasses and mastodon diets were tree vegetation.
And the following ChatGPT map provides information about Paleoindian (Clovis) sites, known, presumed, or inferred to be about 13,000 years BP:
Whereas I take issue with the delineation of both the Laurentide and Cordilleran ice sheet margins on the above map, specifically the suggested open corridor between these ice sheets and as well the configuration of the ice margin in the northern New England area, the above maps are given here simply as a means of underscoring and giving a sense of the relevance and importance of the findings of this VCGI study to the issue of Paleoindian presence and migration.
One of the side benefits of this present VCGI study is that the identification and delineation of ice margins, identifying open, ice-free terrain, and associated water bodies give a sense of possible migration routes and more favorable locations for further search for Paleoindian sites in Vermont. Obviously, the combined input from glacial history with archaeology would seem to offer potential benefits.
I am not an archaeologist and am on speculative thin ice, but I am intrigued by the relationship between glacial history and early human presence in Vermont. My understanding is that two sites in Vermont represent early Paleoindian (Clovis) presence. One is an upland site near Ludlow and the Okemo ski area, referred to as the Jackson Gore site, and the other is a lowland site near Williston, known as the Mahan site. The Jackson Gore site is in the Connecticut Basin, close to the divide with the Champlain Basin. The site has been documented by fluted Clovis spear tip points made of various lithologies, some of which have been identified as coming from Maine and Pennsylvania. Intriguingly, wooly mammoth skeletal remains were found in 1848 at a railroad excavation site about four miles west of the Jackson Gore site, near Healdsville, Vermont. These remains have been C14 dated as 12,800 years BP. 1 Reference an article by Timothy Page, entitled Vermont’s ice-age icon, the Mount Holly Mammoth, https://vermontdailychronicle.com/vermonts-ice-age-icon-the-mount-holly-mammoth/amp/ Thus, the Healdsville and Jackson Gore sites are approximately the same age.
The Mahan site is regarded as the earliest lowland Paleoindian site in Vermont. Fluted Clovis points were also identified at this site, which thus may be approximately the same age as the Jackson Gore site. I understand from the Web that archaeologists speculate that this site may have been a summer base camp where a group of 25–40 individuals stayed for an extended period. It is said to have offered strategic advantages: a slight hill for elevated views, a warmer microclimate, and easy access to water, game, and edible plants. Again, according to the Web, the lithic origin of projectile points at the Mahan site includes local materials (Champlain Valley & Vermont sources, including quartzite and chert from nearby deposits, likely from local river cobbles and bedrock outcrops), and as well non-local, exotic materials:
- Munsungan chert (from northern Maine, (~300–400 miles away)
- Jasper and rhyolite from Pennsylvania (~400–500 miles south)
- Possibly other fine-grained lithics from New York and Québec.
Multiple Paleoindian sites have likewise been identified in Quebec. The Cliche-Rancourt–Mamsalhabika (CRM) archaeological site is the oldest Paleoindian site in Quebec. According to ChatGP, this site reportedly is located on a sandy terrace along the Araignées River in the Lac Megantic region near the Maine border. The CRM site is said to include stone tools and cultural traces dating back ~12,000 years, connected to early human occupation associated with Clovis culture. Projectile points include a mix of local and long-distance provenance raw materials, the main ones being Red Munsungun chert (northern Maine), Rhyolites from northern New England, notably Mount Kineo (Maine) and Mount Jasper–Jefferson (New Hampshire), and local volcanic rocks (rhyolite / trachyte / phonolite) from the Montagne de Marbre area.
In general, it appears that the Jackson Gore, Mahan, and CRM sites are approximately the same age. The estimated Clovis age dates of 12,500 –13,500 BP correspond with late T7 and early Champlain Sea time, using the deglacial chronology and findings developed in this report. This dating is based on a correlation of late T3 time with the White Mountain Morainic System which is dated by Thompson et al at about 13,800 – 14,000 years BP, versus the dating of the incursion of the Champlain Sea by Cronin et al at about 13,000 years B.P. The deglacial history of Vermont is discussed in detail below, but the following is a brief summary intended to give a sense of the human presence in Vermont relative to glacial history.
The following map shows the ice margins and associated ice sheet extent delineated in this study, and as well locates the Jackson Gore and Mahan sites:
The oldest glacial times, as identified in this present study, which are not marked on the above map, are represented by Disconnections in T1 and T2 times in the “Nunatak Phase,” when the ice sheet in the Connecticut Basin became Disconnected from the parent Laurentide ice sheet in the Champlain and Memphremagog basins. In general, subsequent deglacial history is represented by step-down recession of ice margins in a reverse gradient setting from Late T3 to T8 times. The above map shows Late T3 time by the pale blue-green color, T4 time by the blue color, and T6 time by the maroon color. The T5 margin is not shown owning to scale limitations, but was intermediate between the T4 and T6 margin.
The age dates for the Champlain Sea incursion are associated with a portion of T7 time. T7 was a complex time in Vermont. As per the findings of this present study, as discussed in detail below, all ice margins were diachronic. However, temporal differences for the Champlain lobe were especially important in T7 time. The Champlain lobe became destabilized by three distinct, multiple events in T7 time:
- T7-1 time: Beginning in early T7 time, the sudden and substantial lowering of Lake Vermont from the Coveville to the Fort Ann levels was associated with destabilization of the Champlain lobe (as marked by “Streaks,” which are LiDAR features discussed in the Addendum), the transformation of the eastern margin of the lobe, along a “Disaggregated” corridor, from a lateral type of margin toward a frontal type of margin, with calving in narrow basins on the Basin floor. The Lake Fort Ann corridor is not specifically marked on the above map, but corresponds with the outer fringe of the dark green T7 footprint on the above map. As can be seen, the shape of the Champlain lobe at this early T7 and Fort Ann time was long and convex.
- T7-2 time: This was followed by a first phase of ice streaming, in later T7 time, triggered by loss of buttressing support associated with the physiography of the basin floor, as marked by Lateral Shear Margin features (termed “Scarps”), Mega-Scale Lineations, and calving. This streaming occurred primarily in the southern portion of the Champlain lobe, south of an Onset Zone (near Shelburne, just south of Burlington), as marked on the above map. The footprint of T7-2 time was similar to the T7-1 time, with slight recession. Again, the shape of the Champlain lobe still retained its long convex shape at this time.
- T7-3 time: A second phase of streaming was triggered by the sudden and substantial lowering from the Fort Ann level to the Champlain Sea in very late T7 time. The frontal tip of the Champlain lobe at this T7-3 time is marked by erosional features associated with the Onset Zone, descending down from the Fort Ann to the Champlain See level. The configuration of the Champlain lobe at this time is marked by the lemon green color on the above map. This second phase of streaming was accelerated and may have been associated with the collapse(?) of the Champlain lobe, with complete and rapid recession of the Champlain lobe from Vermont, as marked by calving ice margin features.
Whereas the position of the Champlain lobe ice margin at the time of the initial incursion of the Champlain Sea (T7-3 time) has been traditionally and conventionally thought of as being located north of the Quebec border, with a flattened Champlain lobe, the evidence presented in this study suggests that the Champlain lobe remained as a long and convex lobe at this late T7 time, extending southward in the Champlain Basin with its frontal tip again at the Onset Zone near Shelburne. Again, the above map shows the footprint of the Champlain lobe at the beginning of this T7-3 time by the lemon green color. The above map also shows the approximate trace of the Champlain Sea footprint at the marine limit. At this early, initial Champlain Sea time, the Champlain Sea thus occupied a narrow, “Disaggregated” corridor along the lateral eastern margin, and as well an open area south of the lobe tip. Again, the incursion of the Champlain Sea into the Champlain Basin has been dated as at about 13,000 BP, which is in the age range of the Clovis culture. The Champlain Sea persisted while the ice sheet margin rapidly receded from Vermont, followed by a slight readvance back into northernmost Vermont in subsequent T8 time, which is not shown on the above map but is correlated with a readvance in Quebec recently reported by Parent (2023). According to Etienne Brouard (2026, personal communication) the evidence in Quebec may support the concept of Champlain Sea in a narrow corridor along the ice margin in Quebec at an early time.
To give a regional perspective with regard to human presence as just described and potential migration pathways, the following is a regional map of Vermont and neighboring states, specifically New Hampshire, New York, and Maine, and the Province of Quebec. Also identified is Lake Champlain and the St. Lawrence River, the approximate locations of the Jackson Gore, Mahan, and CRM sites, and the position of the very late T7 ice margin associated with the Champlain Sea incursion. Whereas the Champlain Sea is not shown, its footprint at this time would have extended inland from the Atlantic along the southern margin of the ice sheet which at that time was tri-lobed with the St. Lawrence lobe, the Champlain lobe, and the Ontario lobe, all of which is discussed in the following report.
This map shows possible Clovis time migration routes by the large block arrows. Thus, as can be seen multiple ice-free pathways along waterways were available for the three Clovis sites, including possible interchange. These early itinerants likely witnessed the lingering presence of ice in the Champlain lobe, as shown by the above map, but large masses of ice elsewhere may have persisted for a long time. For example, whereas the Jackson Gore site area may have become ice free at T2 time, it is also possible that remnants of the ice sheet persisted for a long time. The Jackson Gore site is located on the eastern side of a col where the ice became Disconnected from the Champlain lobe parent ice sheet, and may have persisted with its insulating sediment cover, which according to the literature tends to be substantial on such deposits.
Similarly, the Mahan site is on the flank of today’s Mud Pond which formed from an ice block, essentially as a kettle hole, which my probings back in the 1970s showed is about 50 feet deep. I collected a bottom sample of peat from this site for C14 dating, but the resultant date proved to be anomalously too young, likely reflecting cross contamination. It is likely that ice persisted in this kettle hole and may have remained in the time of human presence at the Mahan site.
The ice margin positions which are identified in this VCGI report provide more specific information than previously available about possible early human migration pathways into and around Vermont, and deserve further scrutiny in this context. Early human history in Vermont is an intriguing sidebar to this deglacial ice margin recession history, going beyond my expertise and the scope of this report. But it seems obvious that a further dialogue might be productive. From my perspective, hiking, canoeing, or rafting explorations into Vermont as for example via the ice-free Champlain Sea corridor might have been fabulous experiences, with dramatic vistas, though it is likely that Clovis peoples had survival, and not tourism in mind.
Franzi proposes re-naming Coveville Lake Vermont as “Lake Abenaki” and Fort Ann as “Lake Akawasasne” in honor of Native Americans, which in my view is quite appropriate, although the Akawasasne and Abenaki peoples generally are regarded as younger than Paleoindian time, postdating Champlain Sea time.