NEW YORK, JULY 22, 1892.

A VISIT TO A "PICT'S HOUSE."

BY DAVID MAC RITCHIE.

As I have to-day visited an admirable specimen of the underground structures so frequently found in Scotland, where they are popularly known as "Picts' Houses," some description of it will, I think, prove interesting to the readers. of Science, although the place itself has long been known to antiquaries. There are very many examples of these structures in the British Isles, notably in Scotland and Ireland, but unfortunately the information regarding them (almost invariably most exact and detailed) is for the most part buried in the various volumes of "Transactions" of antiquarian societies, and is thereby practically useless. If the descriptions already published regarding these buildings, together with reproductions of the diagrams illustrating them, could be focussed into one volume, the result would be of the highest interest to those who have paid attention to the subject, and would be a positive revelation to those who have not yet done so.1 And one great advantage to be derived from a comparison of the various delineations would be that the student would realize that, although such structures are referred to under many names (such as underground caves, souterrains, weems, cloghauns, Picts' Houses, and — popularly — fairy halls), they all belong to one great

class.

The specimen visited by me to-day is situated at Pitcur, in Forfarshire, about two miles to the south-east of the small town of Coupar-Angus, and is locally known as "the Picts' house." It is entirely beneath the surface of the ground, and the portion of it which is still covered over stretches for about twenty feet beneath a ploughed field. That is to say, its roof is covered by a foot or two of soil, through which the plough passes without ever striking the flat, stone roof below. In other cases, indeed, the ploughshare has often been the first discoverer of these subterranean galleries.

The

The ground-plan of the Pict's House at Pitcur may be roughly described as of a horseshoe shape, with a shorter gallery parallel to the exterior curve of one side. horseshoe itself is about 130 feet in length from end to end, with an average depth of 6 or 7 feet, and an average breadth of about 6 feet. The shorter gallery is about 55 feet long, and its dimensions otherwise resemble those of the horseshoe, except that it broadens out into a bulbshape at the inner end a common feature in such structures. The line of length, in each case, is taken along the middle of the gallery, there being, of course, a great difference between the length of the inner and outer curves.

Be it understood that both of these galleries are, as it were, great symmetrical ditches or drains, quite underground, and entered by several burrow-like doorways. Their sides have

1 I may mention that, as a small beginning in this direction, I am about to issue a pamphlet (published by David Douglas, Edinburgh) containing several written descriptions and sketches of such structures; extracted from the "Proceedings of the Society of Antiquaries of Scotland."

been carefully-built walls of large, unhewn, unmortared stones, and these are still to a great extent unimpaired. The roof was formed by bringing the upper tiers of the wall slightly together, and then placing huge slabs of stone across from side to side. Two of the largest of these roofslabs measure as follows: One (the largest of all) is about 74 inches in length, by 58 inches in breadth, and from 11 to 13 inches in thickness, its shape being an irregular oblong. The other is about 60 inches long, by 48 inches broad, and 12 inches thick. These are certainly very large specimens, but one is always struck by the great size of the flag-stones used in roofing these underground retreats. I have described as unhewn all the stones employed in this building, but (as in similar cases) one is led to conjecture that some rough process of shaping must have been adopted, although the outlines are perfectly rude, and no trace whatever is visible of any tool. The selection of these great stones, whether from a quarry or a hillside, their carriage to the scene of action (often from a very great distance), and the method used in placing them in position, are all problems which have greatly puzzled antiquaries.

In the Pitcur "house" most of the roof-slabs have disappeared, having obviously shared the fate of so many monuments of antiquity, at the hands of proprietors and farmers in need of building materials and quite devoid of all interest in archæology. But (perhaps because it goes underneath arable land) the northern portion of the great horseshoe gallery still retains its roof; and this part of the building is, therefore, in all probability, in its original condition. It appears to have been of itself a "house," apart from the main gallery of which it forms a portion, for it has a carefully-built doorway leading into the main gallery; and, moreover, an extra ascent to the upper earth leads from the side of the wall just at the outside of this doorway. On going through the doorway of this inner portion, one finds, on the right hand, a small recess in the wall, about 33 inches high, 23 inches broad at the floor, and going into the thickness of the wall about 21 inches. Although this cavity is 23 inches broad at the base, the two slabs which form the supports of its little doorway are made to slant towards the top, where the breadth narrows to 14 inches. Within this recess it is possible for a man of 5 feet 10, and of proportionate breadth, to sit in a squatting posture; but it is a very "tight fit." I am particular in giving the dimensions of this recess, because the late Captain Thomas, a naval officer who devoted much time and study to these subterranean structures, and who found this little recess on the right hand of many of their doorways, regarded them as probably identical with the "guard-cells " of the Pictish "brochs." Captain Thomas quite realized that if these were really "guard cells" they were useless for any but men of distinctly small stature an attribute of the Picts, according to tradition.

It is difficult to convey a true idea of such buildings by written description alone, but perhaps these notes will give the readers of Science some impression of an example of a very interesting class of structures.

Easter Logle, Perthshire, Scotland, July 1.

KEY TO THE MAYA HIEROGLYPHS.

BY CYRUS THOMAS, PH.D.

I GIVE here in as limited space as possible a list of the Maya letter 'glyphs so far as I have determined them, together with the corresponding phonetic equivalents; and some examples of my attempts at deciphering the written characters of the Codices.

It is necessary to explain that the letter-equivalent given to each is to be understood as only the chief phonetic element of the character represented, for, in most cases, more than this chief or prominent element is included in the one symbol. The consonant sounds are those chiefly represented, but the character, as a rule, combines therewith a vowel and sometimes even a subordinate consonant sound. happens that the same consonant sound is represented by several different characters depending upon the subordinate phonetic elements combined with it. A change, however, in the character does not necessarily follow from a change in the order of the phonetic elements it represents; thus, what denotes ci as a prefix may stand for ic or c at the end. The examples given of the added vowel and subordinate

k appears to be a combination of Nos. 3 and 5. The latter sometimes contains the dotted portion seen in 6. No. 6 is frequently found where it must be interpreted che, "wood," yet occurs without the dot-surrounded portion where it has the same signification. Other variants are found in the Dresden Codex.

7. K'.- Found as ke and ek, also as Ce.

8. Ch'. Sometimes chi, as in the symbol for Chikin, "west;" 'Ch' as final. Landa's first x appears to be an attempt to give this character which is the partially closed hand.

9. KU'.- Landa's symbol. This does not appear to be subject to any variations that affect its phonetic value.

10. X.-Cross-hatching usually indicates x (sh) as the leading phonetic element; however, it is sometimes rendered by ch', as is evident from its appearance in the symbol for the day Chicchan if we consider it phonetic. However, the day symbols cannot always be relied upon in this respect, as will be seen by what follows.

11. X'.- Landa's second x is substantially the same as this character. But he has taken two characters for one, as in this the x is represented by the dotted lines alone; the

little loop at the forehead, or rather the little parallelogram, in it is a; the face character n. The whole character appears to be properly rendered by xan, "slowly, leisurely, gently." The chief variation in the combination is found in the loop at the forehead, which may be a vowel or consonant. This form of x is seldom found except in combination with n.

12 (a, b, c). E and Ee. - The variations are shown in 126 and 12c.

13 (a, b, c). L'.- This is Landa's first 7. The variations are shown in 13b and 13c. Found in combination with different vowels, as le, ol, etc.

14. L'.-If Landa's second 7 be turned round it will be found to be a rude imitation of this character, which is the symbol for the day Ahau. Li, in the symbol for Likin, East; follows ku, etc.

15. M',- Me.-Symbol for the day Men.

16 (a, b). M'.- Varies in having the little loops at the top, sometimes solid, as in 16b. The dot-surrounded portion of 16b is used alone in one series of the Cortesian Codex for this letter followed by e. The combinations have not been traced.

17. M'. This appears to be another form of m, or m

-

doubled, or combined with n. Not satisfactorily tested as yet, though m is certainly the chief phonetic element.

18 (a, b). M' (?).- Although not thoroughly traced, I am satisfied that this character, which is the symbol of the day Muluc, has m as its chief phonetic element, generally with o or u. The part representing the c is omitted from the day symbol, but is found in the little ring and loops in 18b. The form of the contour of a character is generally of no significance as it may be round, square, or deeply notched without any change in its meaning.

T. Cresson of Philadelphia subsequent to the first notice, in Science, of my discovery, I am much pleased to learn that he has reached a similar determination as to some of these letter symbols by an independent method. As I was not aware until the publication of the article mentioned, that he was at work on the Maya characters, this agreement in our conclusions is highly gratifying, and serves to strengthen both in the conviction that we are making genuine progress in the solution of this difficult problem.

I give here a few interpretations of groups of compound characters to illustrate the combinations of the letter symbols.

Fig. 2 represents a group of four compound characters in the upper division of Pl. XXII* Codex Troano, to be read in this order: upper left, upper right, lower left, lower right; which we will number in the order given 1, 2, 3, 4.

The following is probably a substantially correct translation: (1) U-Zabal, (2) U-le, (3) Cutz, (4) 2-yaxkin: "Set (or literally do the setting of) the snare for the turkey on the second day of Yaxkin." I can give no explanation of the little crosses above the symbol for Yaxkin. The prefix to No. 1 and to No. 2 is the character for u; the upper character in No. 1 appears to be the symbol for z reversed; the band across the lower character the b (possibly interchangeable with p). The figure below agrees very well with this interpretation.

FIG. 2.

19. P or Pp (?).- Although I have not tested this satisfactorily, I am certain from my examinations that its phonetic equivalent is p usually pp. There are some variations found chiefly in the lower portion. The p and b appear to be interchangeable in the Codices even in the same word; for example in the Dresden Codex 48c, we find the b character in the symbol for the month Pop, while on 50b it is replaced in the same month symbol by our No. 19.

20, 21, 22. T".-These characters (20, 21 and 22) appear to have t as their chief phonetic element, varied according to the markings in the upper portion. No. 20 is also varied by the marks in the lower or middle circle.

23. Th'. Is followed by e and i.

30. Cab. The signification of the appendage so often found attached below this symbol has not been ascertained. 31. H.-Sign of aspiration, the open ends always turned toward the character with which it is connected.

32. Kin.-Sometimes without the wing. The latter appears to be used for n, the circle for ki.

33. Kal. If the separate elements are represented, it is probable the section with the dotted line stands for the kand the curved line with the two little teeth for the l.

Having submitted samples of my interpretation to Dr. H.

FIG. 3.

The group shown in Fig. 3 is found in the lower division of plate 26 Cortesian Codex. The characters are taken and numbered in the same order as in Fig. 2. No. 1 is supposed with good reason to be a deity symbol, the name however undetermined. Assuming this to be correct, I translate the group as follows: (Deity) xan yalcab kal-cab, "As" or "in the name of (the deity) slowly gather the swarm of bees and inclose them in a hive."

The figure below shows a priest wearing the mask of the supposed deity hence we say as."

FIG. 4.

Fig. 4 is a group from the middle division of plate XXXII* Codex Troano. The characters are numbered in the same order as the preceding and are translated as follows: Mulcin ku ci- (god of death) xaan; "Collect together for the temple of the holy god of death palm wood." The picture below represents individuals bearing in their hands what appear to be blocks of wood on each of which is the symbol for che "wood."

The little character at the forehead in No. 4 is the symbol

for aa which is found in other combinations where it has the same signification.

So far I have found no marks indicating the plural; this may be represented by duplications.

OSTEOLOGICAL NOTES.

BY D. D. SLADE.

THE jugal arch is present in all of the order Rodentia, and is generally complete, although it exhibits many modifications in its composition. Three bones form the arch, which is straight or slightly curved horizontally, while it almost invariably presents a curvature downwards. The position of the jugal therein serves as a determining character in grouping the various families of the order.

The temporal fossa is often little developed, showing feeble energy in the action of the temporal muscle. On the contrary, the pterygoid plates and fossæ are often largely increased in relation to the enlarged development of the muscular insertions. In close connection with these conditions, the coronoid process of the mandible is small, and even rudimentary, while the parts about the angle are largely expanded. The condyle is little elevated and presents, with few exceptions, an antero-posterior articulating

surface.

Post-orbital processes of the frontals exist in a few of the families, but there is in no case a corresponding process from the arch. The orbit is never separated from the temporal fossa.

In many of the rodents there is present a more or less extensive dilatation of the infra-orbital foramen, through which passes, in addition to the nerve, that portion of the masseter muscle which has its insertion upon the maxilla. This extends around the back of the jugal process of the maxilla in a pulley-like manner, to an insertion just below the socket of the mandibular præmolar, and thus co-operates with the temporal in moving the mandible in a vertical direction. This attachment of a head of the masseter is peculiar to the order, and explains the use of the vacuity in the maxilla which is oftentimes of vast relative proportions.

Assuming the present classification, all existing Rodentia may be brought into two groups, the Simplicidentata and the Duplicidentata. The first embraces the Sciuromorpha, Hystricomorpha, Myomorpha, and the second, the Lagomorpha.

In the Sciuromorpha, the jugal forms the greater part of the arch, extending forward to the lacrymal, and posteriorly to the glenoid cavity, of which it forms the outer wall, and it is not supported below by a continuation backwards of the process of the maxilla. In the more typical forms there is no enlargement of the infra-orbital opening, while the post-orbital processes of the frontals are characteristic of the family Sciuridæ. The external pterygoid plate is entirely wanting, and there is no fossa.

The jugal arch in the Myomorpha is for the most part slender, and the jugal, which does not extend far forward, is supported by the continuation below of the maxillary process. The zygomatic process of the squamosal is short. No post-orbital process of the frontal exists. The infraorbital opening varies. In the family Muridæ, especially in the typical forms, this opening is perpendicular, wide above and narrow below, while the lower root of the zygomatic process of the maxilla is flattened into a thin perpendicular plate. Very much the same condition exists

in the Myoxide, while in the Dipodidæ the foramen is as large as the orbit, rounded, and has a separate canal for the nerve. The malar ascends to the lacrymal in a flattened plate. In close connections with these conditions the coronoid process of the mandible is small and rudimenary, while the parts around the angle of the ramus are much developed.

In the Hystricomorpha the arch is stout. The jugal is not supported by the continuation of the maxillary process, and generally does not advance far forward.. The infraorbital vacuity is large, and is either triangular or oval. The coronary process and the condyle are but slightly elevated above the dental series.

In the Chinchillidæ the jugal extends forward to the lacrymal. In the Dasyproctidæ, Cælogenys is characterized by the extraordinary development of the jugal arch, which presents an enormous vertical curvature, two-thirds of the anterior portion of which, constituting the maxilla, is hollowed out into a cavity which communicates with the mouth. The nerve passes through a separate canal, adjacent to the infra-orbital opening.

In the sub-order Duplicidentata, the jugal arch is well developed. In the family Leporida there are large wing-like, by the maxillary process, continues posteriorly to aid in the post-orbital processes, while the jugal, but feebly supported

formation of the outer side of the glenoid articular surface, passing beneath the process of the squamosal.

In the Lagomyidæ there are no post-orbital processes, and the posterier angle of the jugal is carried backward nearly to the auditory meatus. The infra-orbital opening in the Duplicidentata is of the usual size. The angle of the jaw is rounded and the coronoid process much produced upwards.

In considering the significance of the jugal arch in the Rodentia, the peculiar vertical curvature downwards, which has already been noted, and which is a decided manifestation of weakness, must be taken into account. This condition is compensated in some of the families by the unusual arrangement made in the distribution of the muscular insertions of the masseter through the infra-orbital opening, by which increased energy is imparted to the powers of mastication, and whereby the action of the mandible is rendered fully equal to the demand upon its efforts.

In those families where the above condition does not exist it is evident that the strength of the arch is still sufficient for the antero-posterior movement of the articulation so peculiar to the Rodentia and so characteristic of the act of gnawing.

The relation of the arch to the neighboring parts must also be remarked. For example, the ascending ramus of the mandible differs according to the food. Elevated in the Leporidæ, it is short in the Sciuridæ, and still shorter in the Muridæ.

In the first the coronoid is broad, projects but slightly, is near the condyle, and far distant from the molar series, while the angle of the jaw is broad and well rounded, as in the Lagomyidæ.

In the other two families, squirrels and rats, the coronoid is feeble, pointed, and placed at equal distances between the condyle and the last molar; thus the masseter does not possess a leverage as advantageous as in the hare. This mus cle, however, in the rats has its maxillary attachments much developed, while few fibres spring from the arch—a condition correlative with the feebleness of this last.

Cambridge, June 21.