Search Website Design and Content © by Eric Krause, Krause House Info-Research Solutions (© 1996)
      All Images © Parks Canada Except Where Noted Otherwise
Report/Rapport © Parks Canada / Parcs Canada  --- Report Assembly/Rapport de l'assemblée © Krause House Info-Research Solutions

Researching the Fortress of Louisbourg National Historic Site of Canada
  Recherche sur la Forteresse-de-Louisbourg Lieu historique national du Canada


Preliminary Report on Roofs and Roofing

by Christian Pouyez

[Christopher Moore, Translator]

In Historians,
Preliminary Architectural Studies,
Volume 03, Unpublished Report HG 02
(Fortress of Louisbourg, 1972,
Report Number H G 02 03 01 E)


This study of roofs and roofings is based almost exclusively on the information in the Domestic Architecture File of the Fortress of Louisbourg. Since the file is not yet complete, this report is only provisional in nature, presenting no final conclusion on the structure of roofs.

Three problems will be considered in the course of this study. The first part will consider materials used for roofings of various buildings. The second will try to analyse the most commonly used roof styles. For this section it has been necessary to break the policy of using only the Domestic Architecture File for preliminary reports. We have studied some of the plans of the town of Louisbourg. Finally, in the third part, we will study the structure of roofs, though the available information on this subject is very limited.

The drawings and diagrams accompany this report were made by the draftsmen of the Fortress of Louisbourg from information supplied by the author, who offers his sincere thanks to them.

[PAGE 7:]



The nature of the documents we have used makes information on the roofings of buildings fairly abundant. In the notarized documents - inventories, deeds of sale, agreements, leases, etc. - which constitute the major part of our documentation, it is usual to describe a house by specifying first the materials used for the walls - pickets [piquet], timber [charpente], etc. and then the material of which the roofing is made.

Consequently, it has been possible to determine the type of roofing of 115 private buildings: "cabanes", storehouses, houses, etc. Public buildings such as the barracks, the King's storehouses, and the hospital have been omitted, as examples of different architecture.

Five major types of roofing were used at Louisbourg. Table 1 gives the distribution of the five over five year periods starting in 1715. The dates in this table should be used with care. They are not dates of construction but simply the year in which a given building was mentioned in a document. A house with a bark roof may have been built in 1713 yet not mentioned in a document until perhaps 1732. The dates here only show trends. That almost no references to bark or plan roofings are found after 1739 no doubt means that this type of roofing was only used during the first years of the colony.

[PAGE 8:]



(1) 1715-1719: COVERING: SHINGLE - 0; Covering: "Plan" - 2; Covering: SINGLE BOARDS - 1; Covering: BARK - 2; Covering: DOUBLE BOARDS - 0; Total - 5

(2) 1720-1724: Covering: SHINGLE - 2; Covering: "Plan" - 7; Covering: SINGLE BOARDS - 2; Covering: BARK - 0; Covering: DOUBLE BOARDS - 0; Total - 11

(3) 1725-1729: Covering: SHINGLE - 1; Covering: "PLAN" - 4; Covering: SINGLE BOARDS - 4; Covering: BARK - 2; Covering: DOUBLE BOARDS - 0; Total - 11

(4) 1730-1734: Covering: SHINGLE - 4; Covering: "Plan" - 2; Covering: SINGLE BOARDS - 5; Covering: BARK - 6; Covering: DOUBLE BOARDS - 0; Total - 17

(5) 1735-1739: Covering: SHINGLE - 7; Covering: "PLAN" - 11; Covering: SINGLE BOARDS - 5; Covering: BARK - 2; Covering: DOUBLE BOARDS - 3; Total - 28

(6) 1740-1744: Covering: SHINGLE - 3; Covering: "PLAN" - 1; Covering: SINGLE BOARDS - 3; Covering: BARK - 0; Covering: DOUBLE BOARDS - 0; Total - 7

(7) 1745-1749: Covering: SHINGLE - 0; Covering de "PLAN" - 0; Covering: SINGLE BOARDS - 1; Covering: BARK - 0; Covering: DOUBLE BOARDS - 0; Total - 1

(8) 1750-1754: Covering: SHINGLE - 13; Covering: "PLAN" - 0; Covering: SINGLE BOARDS - 5; Covering: BARK - 0; Covering: DOUBLE BOARDS - 0; Total - 18

(9) 1755-1759: Covering: SHINGLE - 14; Covering "PLAN" - 1; Covering: SINGLE BOARDS - 2; Covering: BARK - 0; Covering: DOUBLE BOARDS - 1; Total - 18

(10) TOTAL: Covering: SHINGLE - 44; Covering: "PLAN" - 28; Covering: SINGLE BOARDS - 28; Covering: BARK - 12; Covering: DOUBLE BOARDS - 4; Total - 116

(11) TOTAL %: Covering: SHINGLE - 37. 9; Covering: "PLAN" - 24.1; Covering: SINGLE BOARDS - 24.1; Covering: BARK - 10.4; Covering: DOUBLE BOARDS - 3.5

(12) TOTAL - 100

The most frequently used roofings were shingles (37.9%), plan (24.1%) and single boards (24.1%). Bark and double boards were much more rare. It will be noted that private buildings were not roofed with slate, no doubt because of its cost and the difficulties of obtaining it. Only certain public buildings had slate roofing. [PAGE 9:]


Shingles were by far the most common roofing material and they are the type on which the most information can be found.

Shingles were used above all for residences, but they were also found on storehouses, sheds and stables (TABLE 2).



(1) Use of Building: - RESIDENCES: Number = 35; Percentage = 79.6%

(2) Use of Building: - STOREHOUSES: Number 6, Percentage = 13.6%

(3) Use of Building - SHACKS AND STABLES: Number = 3; Percentage = 6.8%

(4) Use of Building = Total 44; Percentage = 100

It seems that this material was more often used on fairly substantial buildings such as stone or timber [charpente] houses than on picket [piquet] buildings (TABLE 3).



(1) Construction type: TIMBER [CHARPENTE] HOUSE - Number = 18; Percentage = 40.9

(2) Construction type: PICKET [PIQUET] BUILDING - Number = 9; Percentage = 20.5

(3) Construction type: STONE HOUSE - Number = 2; Percentage 4.5%

(4) Construction type: UNKNOWN - Number = 15; Percentage = 34.1%

(5) Construction type TOTAL Number = 44; Percentage = 100

However there are too many cases where the type of construction is not known for a definite conclusion to be drawn.

Contrary to the ordinary usage in France at that time, the inhabitants of Louisbourg did not place shingles directly on to simple laths (or roofing strips) nailed to the rafters, but on a covering of boards. The difference is easily explained by the Louisbourg climate. The violence of the wind during blizzards made necessary the strongest, most waterproof roofs possible.

[PAGE 10:]

Unfortunately only a very few documents specify the type of board and the type of joint connecting, them. In 1752 Sieur Dubenca agreed to build Sieur Dugué a "charpente" house of which

the roofing will be made of Boston boards with a one "pouce" bevel "à sifflet", covered with shingles [NOTE 1]. [underlining our emphasis]

In another case, it was specified that the roofing would be

of good Boston boards with bevels of two or three "pouces" on each one nailed with two nails on each rafter. The shingles placed on these boards are to have at least four pouces of weather [NOTE 2]. [underlining our emphasis]

[PAGE 11:]

That means that a close fit along the intersection of two adjoining boards was ensured by the bevelling of each board to permit joints "à sifflet". It must be understood that the term "à sifflet" did not have the same meaning at Louisbourg, as in France. In France a joint "à sifflet" was an end joint, formed by the meeting of two obliquely-cut board ends [NOTE 3]. It seems that at Louisbourg the term referred to a bevelled joint. Perhaps we should explain what the bevelling of a board is before we proceed.

To bevel "délarder" means to remove the wood from the corner of one side of a board. It is usually done with a plane. By this method one gets boards with a bevel of one pouce, two pouces or more, so that the planks fit together (figure 1) [PAGE 12: Figure 1 presently unavailable]. A text of 1738, explains clearly how and why this type of joint was used:


... Boston boards which are used for the roofing of buildings. Note that a three pouce bevel was planed into them to make them a better match so that they fit together better and prevent the snow from getting in during the great blizzards. In this way one avoids having to use "torchis" [a mixture of straw and mud] which for one thing is too heavy and for another crumbles when it dries [NOTE 4].

This joint along the length of the board was described in different ways in eighteenth century documents. Usually they simply say that the boards overlapped "à sifflet" [NOTE 5]. Sometimes they mention boards or planks bevelled "à sifflet" [NOTE 6] and specify the size of the bevelled area. Other [PAGE 13:] times they say the boards overlap "en amortissement" [NOTE 7] or that they are bevelled "en chanfrein" [NOTE 8]. All these expressions refer to the same type of joint. Possibly the only difference was in the area bevelled. An "amortissement" of four "pouces" is certainly larger than a bevel "en chanfrein". Before leaving this topic, we should mention that the section of roofing found on lot k of Block 2 (Guion house) included a bevelled board: the shingles are placed on a base of boards with a bevel of 1.5 inches. The boards are roughly 1.4 feet wide [NOTE 9].

The sources we have consulted about shingles have given very little information. Below is found what information we have at the moment. It is taken from three documents and the archaeological analysis of the section of roofing from the Guion house [NOTE 10].


(a) 12 to 13 pouces

(b) 1.40 feet (about 16 pouces)


(a) 4 to 6 pouces

(b) about 6 inches (about 5.5 pouces)

(c) 4 by 5 pouces


(a) 4 pouces

(b) 0.40 feet (4.5 pouces)

(c) 1/3 of their length

(d) 4 pouces

Several problems remain which we have not been able to solve. The thickness of the shingle is never specified for private buildings. The shingles of the roofing found at the Guion house were obviously worn and cannot furnish precise data. How the shingles were fastened is not clear. [PAGE 14] Only one document specifies the use of two nails for each shingle. The only information we have on manufacturing methods, types of wood used and solutions with which shingles were treated comes from the article already mentioned, written by Richard Cox. These are his conclusions:

An examination of over 200 whole and fragmentary specimens from excavations suggested that wooden shingles used at Louisbourg during the eighteenth century tended to be:

(i) Rectangular except for occasional diagonal trimming, of both corners of feather end. (little overall taper in these latter cases).

(ii) Approximately 1.40' x .50' x .03' (butt) and .01' (feather) where taper present

(iii) Manufactured by splitting and slight planing

(iv) Made of pine, cedar or occasionally balsam fir. 

(v) Fastened by iron nails

(vi) Fastened to bevel jointed sheathing of width equal to length of shingle

(vii) Lapped such that 1/3 of the length was exposed to the weather

(viii) Split from saw cut stock such that neither butt nor feather was exactly at right angles to the longitudinal axis of the shingle

(ix) Unpainted (Hanson suggests otherwise for samples from Engineer's House Garden Pool)" [NOTE 11].

[PAGE 15]


Though frequently used during the first years of the colony, "plan" roofings tended to be displaced by shingles and board roofings in later years. The first problem is to discover just what was meant by "plan roofing".

Documents from the beginning of the 18th century often use the expression "plan roofing" without further detail, but is sometimes specified that the roofing was "wooden plan" or "earthen plan" or "plan of wood and earth". Hence it is necessary to differentiate between at least two different kinds of plan, wooden and earthen. That is all that the documents tell about the nature of plan. The reference books we have consulted (Furetière, Trévoux, Diderot, Bescherelle) provide no information. Only the Larousse du XX e siècle gives any suggestion: "plan" could be a corruption of "pelan" which means, "bark peeled off in large pieces for craft uses". By this definition wooden "plan" would simply be bark: "plan" roofings and bark roofings would be identical. This explanation does not seem to be useful in 18th century Louisbourg for two reasons. If the two terms were identical, a document might say either plan or "bark", but not both. Yet there are documents discussing "plan" roofings and bark roofings at the same time, and they seem to mean different things. Second, wooden "plan", as will be seen below, was often used in conjunction with earthen "plan", the earthen "plan" being laid on top of the wooden "plan". That is difficult to imagine if wooden "plan" was really bark. It seems to us, therefore, that at Louisbourg wooden "plan" meant "slabs", that is, the first and the last boards produced when a tree trunk is cut into boards. [PAGE 16:] A slab is flat like a board on one side but curved and retaining its bark on the other. There is nothing to prevent this material being used for roofing. At the moment it seems the likeliest hypothesis.

The reference works have no definition for earthen "plan", but here the documents are slightly more specific and the plans and views permit further precision. The documents have three different phrases which apparently refer to the same thing: earthen "plan" is the most common, but there are also references to turf roofing and sod roofing. It seems almost certain that these three expressions denote what we call today in North America, "sod", that is grass and plants and the earth around the roots. Plans 725-8 and 740-1 give a few examples of houses covered with earthen "plan".

If wooden "plan" were sometimes used by itself for the roof of a building, it was not similar to earthen "plan" which had to have a solid base. Most often earthen "plan" was spread on top of a roofing of wooden plan. Table 4 gives the distribution of the various types of "plan" among buildings with different functions. Two facts should be noted. First, "plan" was




(2) Use - STOREHOUSE: Roofing: "Plan" = 0; ROOFING: Wooden "PLAN" = 1; ROOFING: EARTHEN "PLAN" = 0; ROOFING: EARTHEN AND WOODEN "PLAN" = 0; Total = 1 (4%)

(3) Use: TOTAL = ROOFING: "PLAN = 2 (7%); ROOFING: WOODEN PLAN = 18 (64%); ROOFING: EARTHEN "PLAN" = 2 (7%); ROOFING: EARTHEN AND WOODEN "PLAN" = 6 (22%); Total 28 (100)

[PAGE 17:]

used almost exclusively for houses (96% of the cases studied). Second, there are only two examples where the type of "plan" is not specified and two examples where earthen "plan" alone is cited as the roofing material. That is no doubt due to a lack of precision in the documents. It is likely that the two houses said to have been covered with earthen "plan" actually should be classed under roofings of wooden and earthen "plan". "Plan" roofings were only used for more rudimentary structures. Of 28 cases studied, 21 were picket [piquet] houses. The construction of the other seven houses is unknown, but there is every reason to assume that they too were picket [piquet] buildings.


Bark roofings were fairly common in the first years of the colony, presumably because of the ease with which one could obtain bark. It quickly became apparent that this practice was damaging the local trees, and the king banned bark roofings in 1717.

His Majesty has been informed that it is the habit in Isle Royale to make roofings for houses from tree bark. This can only be prejudicial to the colony, for it will completely ruin the trees. Moreover the fragility of this roofing will make necessary their continual repair and so this roofing will cost as much and more than a covering of boards. Hence His Majesty very expressly forbids the use of tree bark for house roofings and he wishes that, while waiting for slate to become common on Isle Royale, the inhabitants would use shingles for their roofing. [PAGE 18:] He recommends that Sieurs de Costebelle and Soubras make very sure that this is done [NOTE 12].

If the royal decree had been enforced, bark roofings would have been replaced as they rotted away by boards and shingles or another authorized material. It is clear that the ban was not respected by all the inhabitants, for there were bark roofings until 1736. However most inhabitants obeyed the order for the 115 buildings studied include only a dozen with bark roofings, of which five were located out of the town itself: west of the barrachois de Lasson, at the north end of the harbour, and even on the Iles Michaux.

In any case, bark roofings did exist, and not only at Louisbourg. There is reference to one in 1756, on Isle St. Jean. Oddly enough, it was on a building belonging to the King [NOTE 13]. In all the cases we have studied, bark roofings were used exclusively for picket [piquet] buildings, usually houses (table 5).



(1) Use - RESIDENCE: Number = 10; Percentage = 84

(2) Use - STOREHOUSE: Number = 1; Percentage = 8

(3) Use - LEAN-TO: Number = 1; Percentage = 8%

(4) Use - TOTAL = 12; Percentage = 100%

[PAGE 19:]

We do not know what sort of bark was used, nor how and to what it was attached. It seems wise to avoid theorizing on these questions until we are better informed.


This was a common roofing, especially after 1730. It was used for all types of buildings (table 6) and with all kinds of construction materials (table 7). The documents never specify what type of joint was used between the boards in a single layer board roof. Since they were not



(1) USE - RESIDENCE: Number = 17; Percentage = 60.7

(2) Use - STOREHOUSE: Number 7; Percentage = 25.0

(3) Use - SHACK, LEAN-TO: Number 4; Percentage = 14.3

(4) Use - Total: Number = 28; Percentage = 100

covered with shingles, it is likely that they had the tightest joints possible, which would mean a very pronounced bevel joint, or a tongue and groove joint, or possibly a halved joint.

The absence of information on the structure of beams and supports prevents us from knowing whether the boards were placed vertically or horizontally and the maps and plans do not give any help here. There are illustrations of both horizontal and vertical alignments, but it [PAGE 20:] cannot be determined whether they show single or double layer board roofs. Hence no conclusion can be reached.



(1) Construction - PICKETS [PIQUETS]: Number 16; Percentage = 57.2

(2) Construction - CHARPENTE: Number 6; Percentage 21.5

(3) Construction - STONE (BRICK): Number = 1; Percentage 3.5

(4) Construction - BOARDS: Number 1; Percentage 3.5

(5) Construction - UNKNOWN: Number = 4; Percentage 14.3

(6) Construction - Total: Number = 28; Percentage = 100


There are only four examples of this type of roof, on 3 houses and 1 barn. Two of these were picket [piquet] buildings, one was a timber [charpente] house and the construction materials for the other is unknown.

Only one document gives any detail on this type of roofing. When the house of André Monier dit Surgere was rented to François Lucas, the latter agreed to build a picket [piquet] house of 36 pieds by 16 pieds to be roofed with a double layer of boards laid lengthwise [NOTE 14]. Knowing that the roof was built with trusses, a ridge-beam and rafters, but without purlins, we are able to deduce that the first layer of boards was laid lengthwise parallel to the ridge-beam. The second layer of boards could have been [PAGE 21:] laid in the same alignment overlapping the joints of the first layer, or they could have been placed perpendicular to the bottom layer.


Table 8 contains all the available statistics on roofings. To facilitate comparisons, we have included both numbers (in the upper part of the table) and percentages (in the lower part). With this table, one can quickly judge the relative importance of each type of roofing for each building according to its use and type of construction.

Though the table is designed to answer precise questions (e.g., how many shingle-roofed charpente houses were there?), one can draw from it, if not conclusions, at least a few general indications:

(1) the most common roofing for picket [piquet] houses was "plan", while "charpente" houses almost always had shingles

(2) storehouses, usually made of pickets [piquets], had single layer board roofings in a majority of cases. However there were a sizeable number with shingle roofs

(3) references to sheds, cow-sheds, horse stables, barns and other buildings of that sort are too scarce to permit firm conclusions.

The available figures seem to suggest a greater frequency of single layer board roofs, but this is only an indication, to be used with caution.

[PAGE 22:]



Table 8 is presently unavailable.



Information on roof design is of two kinds: that which can be found in the written documents of the period and that which appears on the maps and plans. At the present state of research, the former is extremely rare, and the latter is very difficult to interpret, as an example will show. In the first plans of Louisbourg, 1717-2, 1720-2 and 1720-4, all the buildings were hipped at both ends, except for 5 adjoining houses on Block 4. The most elementary logic will suggest that the mapmaker had simplified the details. It was his task to show where the buildings and houses were, not to reproduce the design of the roofs. Consequently these plans are of no use whatsoever for this study. However not all the plans are as uniform as those. We have found two which seem to be relatively precise in all details: plans 1730-2 and 1734-4. The latter is not complete, unfortunately, but the sample of buildings appearing on it is surely representative of the town as a whole: this can be shown by a comparison of the two plans. Other plans besides 1730-2 and 1734-4 were accurate enough to be used. However information taken from a plan is not useful in this case unless the plan covers a large part of the town, or better still, the whole of it. That eliminates many plans, such as 1725-8, 1731-3, 1740-1, etc. Moreover it is obvious that only the clearest plans on which roof designs can be determined, are helpful. That eliminates several more, particularly N.D. 24. [PAGE 24:]

Hence the material for this chapter comes solely from written manuscripts and the two chosen plans: 1730-2 and 1734-4.

Documentary information is very scarce. We have only been able to identify, and not always precisely, eighteen buildings: two storehouses and 16 private houses. Six variations of roof design can be distinguished. Table 9 gives their distribution.



(1) ROOF DESIGN: 1 GABLE, 1 HIP; Number = 2; Percentage = 11.1

(2) ROOF DESIGN: 2 GABLES [NOTE: Including one mansard roof: the Vallée; Number = 3; Percentage = 16.7

(3) ROOF DESIGN: 2 HIPS; Number = 2; Percentage = 11.1

(4) ROOF DESIGN: 1 GABLE, 1 ONE ADJOINING END; Number = 2; Percentage = 11.1

(5) ROOF DESIGN: 1 GABLE, 1 OTHER END UNKNOWN; Number 8; Percentage = 44.4

(6) ROOF DESIGN: 1 HIP, 1 OTHER END UNKNOWN; Number = 1; Percentage = 5.6

(7) ROOF DESIGN: Total; Number = 18; Percentage = 100

Such a limited series of figures cannot provide anything conclusive: the use of plans and views is indispensable. Table ten gives both sketches and totals of the various roof designs as shown on plans 1730-2 and 1734-4. The table needs no long explanation but an aspect of the methodology should be noted. A comparison of tables 9 and 10 shows the dangers of drawing


[PAGE 25:]



Table 10 is presently unavailable.

[PAGE 26:]



Table 10 (2nd part) is presently unavailable.

[PAGE 27:]

conclusions from incomplete documentation. Though the documents describe only four kinds of roofs, plan 1730-2 shows 16 and 1734-4, though incomplete, shows 14.

In this table, the roof design of military buildings, buildings of lot 1, and the hospital have been omitted. Except for one or two cases, the figures in the table cover only civil architecture.

[PAGE 28:]



Documentation on roofing structures is even more fragmentary than what was available on roof design, and the maps and plans have not been very helpful. The plans in the archives are almost exclusively military buildings and have been eliminated for that reason. When the Domestic Architecture File is complete, we will certainly have available a much larger amount of data. For a number of reasons beyond our control, the file will not be complete for several months. Therefore this part of the report, even more than the earlier sections must be considered provisional and incomplete.

Because the information is so limited, we have studied each case separately, giving the description of structure which appears in the documents, or as much of the description as can be drawn from them. Also provided are sketches illustrating the structure as precisely as possible. Except for those of the Vallée house (case #3) and the Dugué house (case #5) none of these plans are in the documents. They have been sketched from the descriptions - often incomplete - which are in the source materials.

(A) Case #1 - "Cabane" of Laurent Dybarart - 1721 [NOTE 15]

(i) Location: Block 5, lot A

(ii) Dimensions: 32 pieds by 18 pieds

(iii) Type of construction: Pickets [Piquets]

(iv) Roofing: wooden and earthen "plan"

[PAGE 29:]

(v) Roof design: hipped at both ends

(vi) Chimneys: One, of stone and earth up to the level of the attic floor, and of earthen "torchis" (straw packed in mud) above

(vii) Walls and Buttresses: 224 pickets [piquets] about 6 pouces in diameter

(viii) Floors, upstairs and downstairs: rafting timbers of about 4 "pouces" in diameter. Timbers of the lower floor squared on one side only

(ix) Plates: 10 rafting timbers (these were laid end to end along the top of the pickets [piquets])

(x) Structure of the roof: 1 ridge beam; 4 hip- rafters; 6 king-posts; 6 cross-tie beams; rafters


(1) Strange as it may seem, it appears that the 6 cross-tie beams also served as joists to support the floor of the attic. These beams, if we can judge by their price, must have been timbers of very high calibre. They are estimated to be worth 39 "livres", though the ridge [PAGE 30:] NOTE: Figure 2 is presently unavailable. [PAGE 31:] beam, the four corner-rafters, and the six king-posts were estimated to be worth 5 livres 8 sols as a group.

(2) Beside the six cross-tie beams, the document mentions 3 sleepers. Their function is not specified but it is almost certain that they made up the frame for the chimney and the staircase (see figure 2).

(3) The roof had six trusses. There were no common rafters.

(B) Case #2 - Widow Degoutin house - 1724 [NOTE 16]

(i) Location: Block 14 lot G

(ii) Dimensions: 30 pieds long, measured on the outside, 20 wide measured inside

(iii) Type of construction: Charpente

(iv) Roofing: Unknown

(v) Type of Roof: probably gabled at both ends

Apparently the roof of this house was destroyed in a windstorm. The experts said they found "the said timbers in one piece and in very poor condition" and they examined the timbers. They mention amongst other things, four posts. It may be that these were wall posts, torn away when the roof fell, but we find that unlikely for two reasons. First, the experts said they found the frame "in one piece". That would mean either that the whole house fell down, in which case more than four posts would have been found, or that the timbers in question were parts of the roof, in which case the four posts would actually be cross tie beams. Second, the listing of the timbers is done on the proper order beginning with the plates, the posts, the ridge beam, and the trusses. In other words, the

[PAGE 32:]

NOTE: Figure 3 is presently unavailable.

[PAGE 33:]

pieces are named in the order of their positions in the structure of the roof. If the four posts were wall posts, it is difficult to see why they would be listed between the plates and the ridge beam.

From the meager data in the document we can assume that the roof had the following parts (figure 3).

(i) 4 cross-tie beams ("posts")

(ii) 2 trusses with a collar beam, king post and braces.

The principal rafters of the two trusses served as common rafters. Because there were only two trusses for this 30 "pied" building, we assume that the house was gabled at both ends. In that case, the trusses would be spaced 10 pieds apart.

(C) Case #3 - Vallée House - 1752 [NOTE 17]

(i) Location: Block 34, lot C

(ii) Dimensions: 40 pieds by 34 pieds

(iii) Type of construction: Stone. In 1752 both gables were remade with timber [charpente] and stone fill

(iv) Roofing: Unknown

(v) Roof Design: a mansard roof and two gables

The house had twelve rooms, 4 on the ground floor, 4 on the second floor and four in the attic. Each of the eight rooms on the first two floors had a chimney.

Gable and Truss Structures (Figures 4 and 5; plan 1752-12).

[PAGE 34:]

NOTE: Figure 4 is presently unavailable.

[PAGE 35:]

NOTE: Figure 5 is presently unavailable.

[PAGE 36:]

Table A: 2 cross-tie beams ("struts")

(i) Length: 34 "pieds"

(ii) Size 7 x 9 "pouces"

B: 1 cross-tie beam ("strut")

(i) Length 30 "pieds"

(ii) Size 7 x 9 pouces

C: 7 Posts separated by one of the struts 34 "pieds" long

(i) Length 16 "pieds" (ii) Size 7 x 9 pouces

D: 4 posts (i) Length 6 "pieds" Size 6 x 7 pouces

TRUSSES: E: 1 collar beam

(i) Length 14 pieds (ii) Size 6 x 7 pouces

F: 1 King-post

(i) Length 4 "pieds"

(ii) Size 6 x 7 "pouces"


(i) Length 18 "pieds"

(ii) Size unknown


Note the absence of a ridge-beam, and the simple structure of the roof which has no purlins, and where the principal rafters are used as common rafters. In fact we doubt that such a major building actually had no ridge beam. In our opinion, there was a ridge beam, though it is not shown on the plan (1752-12). Perhaps it was forgotten by the draftsman, or more likely the plan was intended to illustrate only the structure of the gables and the timbers necessary to their construction. It is probable that when the gables were rebuilt, the ridge beam was left untouched and consequently was omitted from the plan.

[PAGE 37:]

(D) Case #4 "Cabane" of Andre Monier, dit Surgere - 1736 [NOTE 18]

(i) Location: On the north side of Louisbourg harbour

(ii) Dimensions: 36 pieds x 16 pieds, 7 pieds high under the cross beams

(iii) Type of construction: Pickets [Piquets]

(iv) Roofing: double boards laid horizontally

The house had a double chimney made of stone and clayey earth. There were no floors above or below and there is no reference to plates. In fact, the only information we have is:

... there will be only seven cross beams laid on the said pickets [piquets] without other structures except trusses to hold the ridge beam, and rafters...

We may assume that the roof was made of seven simple trusses whose principal rafters were also common rafters. What are called "traverses" in the document are actually "tirants", cross beams. Possibly there were no plates. The ends of the principal rafters could have rested on the cross beams which themselves lay directly on the posts. Alternately, it is possible that there were plates, which were not judged worth mentioning in the document.

The roofing had the following structural parts (figure 6):

(A) Cross-tie beam ("traverses")

(B) King Post

(C) Ridge-beam

(D) Rafters; [PAGE 38:]

NOTE: Figure 6 is presently unavailable.

[PAGE 39:]

(E) Case #5 - Dugué house - 1752 [NOTE 19]

(i) Location: Block 36, lot C

(ii) Dimensions: L-shaped house, 21 "pieds" wide, 44 "pieds" long on one leg, 45 "pieds" on the other. Two floors and an attic

(iii) Type of construction: "Charpente" (iv) Roofing: Shingles

(v) Roof design: 1 gable, one inverted hip

(vi) Wall plates: 8 "pouces" x 9 "pouces"

(vii) Master posts (4): 10 "pouces" x 10 "pouces"

(viii) Posts where interior partition intersects the walls: 8 x 8 "pouces"

(ix) Other Posts: 5 "pouces" x 8 "pouces" (x) Struts: 8 "pouces" x 8 "pouces"

(xi) Cross beams: spaced every 2 "pieds", 7 x 9 "pouces"

(xii) Plates: 8 x 9 "pouces".

ROOF STRUCTURE: The trusses were spaced ten "pieds" apart. They were made of (figure 7)

(A) Principal rafters, 6 x 7 "pouces"

(B) Collar beam, 6 x 7 "pouces"

(C) King-post, 6 x 7 "pouces"

(D) Ridge beam, 5 x 6 "pouces"

(E) Purlins, 5 x 6 "pouces".

The rafters (F) were 4 "pouces" by 4 "pouces".

(F) Case 6 - Jacques Rabasse house - 1756 [NOTE 20]

(i) Location: near the barrachois crossing

[PAGE 40:]

NOTE: Figure 7 is presently unavailable.

[PAGE 41:]

NOTE: Figure 8 is presently unavailable.

[PAGE 42:]

(ii) Dimensions: 30 "pieds" x 22 "pieds"

(iii) Type of construction: squared pickets [piquets]

(iv) Roofing: "Plan" and earthen "plan"

(v) Roof design: Unknown.

The house was "floored upstairs and down" and had a double chimney of clayey earth.

Evidently the roof had no trusses. The only information in the document is that the roof "is made of round firs". Since the document is a fairly detailed description of the Rabasse house, one might think that if the roof structure had been more complex, there would have been reference to its various parts (ridge-beam, collar beam, king post, etc ... ). As this is not the case, we conclude that it was a simple roof without trusses. The spacing of the rafters is unknown (figure 8).

The documentary material on which this report lies is not comprehensive enough to support definite conclusions, particularly regarding roofing, structures. Hence the few results we have are presented without extensive comment.

It may be noted that we have not raised the question of fittings and decorations for roofs: as yet the sources have not provided evidence of whether or not finials were commonly used, whether the roofs hung over the walls, etc. For information on the problem of the joints used for gutter tiles and chimneys, Henri-Paul Thibault's report on stoves and chimneys is recommended.


[NOTE 1:] Convention faite entre le Sieur Dugué et Michel Dubenca. Louisbourg, 1 octobre 1752. A.N., Section Outre-Mer, G 3, carton 2041, pt. 1, no 131. [NOTE 2:] Devis des ouvrages pour le bâtiment des Srs Duperrier et Rodrigue. Louisbourg, 1737. A.N., Section Outre-Mer, G2, 184, fol. 392-394. [NOTE 3:] P. Chabat, Dictionnaire des termes employés dans la construction. (Paris, Morel, 1876), article SIFFLET. [NOTE 4:] Mémoire au sujet de 1'Isle Royale, s.l., [1738]. A.N., Section Outre- Mer, D.F.C., no 141. [NOTE 5:] (1) Marché entre Mr Beaubassin et Dubenca. Louisbourg, 30 mai 1756. A.N., Section Outre-Mer, G3, carton 2044, no 53. (2) Réparations dans une maison. Louisbourg, 7 août 1752. A.N., Section Outre- mer, G2 vol, 208, dossier 475. [NOTE 6:] Marché entre Mr Beaubassin et Dubenca. Louisbourg, 30 mai 1756. A.N., Section Outre- Mer, G3, carton 2044, no. 53. [NOTE 7:] Ainsi, dans un texte de 1737, on précise que les planches "se recouvriront les unes sur les autres par un amortissement de quatre pouces pour ne faire qu'un plan uny et les planches bien arrêtées et clouées sur les chevrons. "(Marché pour les fortifications de la ville de Louisbourg. Paris,10 mai 1737. A.N., Col., C1lB, vol. 19, fol. 176-188v.) [NOTE 8:] "Devis et conditions à observer par ceux qui entreprendront les ouvrages ordonné par le Roy". s.l., 25 septembre 1753. Archives du Séminaire de Québec, Papiers Surlaville. [PAGE 46:] [NOTE 9:] Richard E. Cox, "Wooden shingles from the Fortress of Louisbourg", Bulletin of the Association for Preservation Technology, vol. II, nos 1-2, 1970, pp. 65-69. [NOTE 10:] Les informations données ici proviennent des sources suivantes, dans l'ordre: (a) "Devis et conditions à observer ...". s.l., 25 septembre 1753. Archives du Séminaire de Québec, Papiers Surlaville. (b) Maison Guion, ilôt 2, terrain K. Réf: Richard E. Cox, op. cit., Bulletin of the Association For Preservation Technology", vol. II, nos 1-2, 1970, p. 65-69. (c) Marché pour les fortifications de la ville de Louisbourg. Paris, 10 mai 1737. A. N., Col., Cl1B, vol. 19, fol. 176-188v. (d) Devis des ouvrages pour le bâtiment... Louisbourg, 1737. A. N., Section Outre-Mer, G2, vol. 184, fol. 392-394. [NOTE 11:] Richard E. Cox, op. cit., p. 69. [NOTE 12:] "Mémoire du Roy au Sr de Costebelle, gouverneur, et au Sr de Soubras Comme Ordoneur à 1'Isle Royalle". s.l., 29 juin 1717. A.N., Col., B, vol. 39, fol. 294v. [NOTE 13:] "Au nonme Charles Pitse, habitant de l'isle st Jean, pour douze douzaine d'écorces qu'il a fourni pour couvrir les forges du Roy.... 18". ("Bordereau des payments qui ont été faits à la colonie de l'isle Royalle.... ". s.l., 20 décembre 1756. A.N., Col., C1lB, vol. 36, fol. 212). [NOTE 14:] Bail à loyer: André Monier dit Surgere, à Francois Lucas. Louisbourg, 19 janvier 1736. A.N., Section Outre-mer, G 3, carton 2039 suite, no 44. [PAGE 47:] [NOTE 15:] Vente de maison: Laurent Dybarart à Michel Daccarette. Louisbourg, 21 novembre 1721. A.N., Section Outre-Mer, G 3, carton 2057, nos 23-24. [NOTE 16:] Visite d'experts à la maison de la veuve Degoutin. Louisbourg, 8 janvier 1724. A.N., Section Outre-Mer, G2 vol. 178 f. 369-370. [NOTE 17:] Diverses pièces concernent directement ou indirectement la maison Vallée: (1) "Conditions de la vente judiciaire de la maison de Sieur Francois Vallée...". Louisbourg, 4 octobre 1732. A.N., Section Outre-Mer, G2. vol. 181, f. 505-547. (2) Vente de maison: Francois Vallée à Joseph Brisson. Louisbourg, 3 décembre 1732. A.N., Section Outre-Mer, G3, carton 2038 pt. 1, no 78. (3) Procès entre Vallée et Brisson. Louisbourg, 27 juin 1733. A.N. , Section Outre-Mer, G 2, v.182, f. 1051-1065. (4) Visite d'experts à la maison Brisson. Louisbourg, 29 juillet 1752. A.N., Section Outre-mer, G 2, vol. 201, no 250. [NOTE 18:] Bail à loyer: André Monier dit Surgere à Francois Lucas. Louisbourg, 19 janvier 1736. A.N., Section Outre-Mer, G 3, carton 2039 suite, no 44. [NOTE 19:] Conventions entre le sieur Dugué et Michel Dubenca. Louisbourg, 1 octobre 1752. A.N., Section Outre-Mer, G3, carton 2041, pt.1, no 131. [NOTE 20:] Vente d'une maison: Jacques Rabasse à dame Madeleine Lartigue. Louisbourg, 30 juin 1756. A.N., Section Outre- Mer, G 3 carton 2044, no 59. [NOTE 21:] H.P. Thibault, "Heating and cooking facilities in private dwellings in Louisbourg (1713-1758)", Preliminary Architect Studies, p. 12.