swelling \ j hi RNSat ii Ma i vie ara : ty y uel “it Wirt orn j ch ii aN 4an# ye ie io Hie ae : PART in tach f Wa caripa tt ait hy Fj Library ad 2 K mate i PROCEEDINGS OF ‘THE ROYAL [IRISH ACADEMY VOLUME XXXIII DUBLIN: HODGES, FIGGIS, & CO., Lrp LONDON: WILLIAMS & NORGATE 1916-1917 ’ ip i: ie a : uy , > vt ces ve ans ik wu oa HF EE MMOL @ 1 ee ALPE) 4 ant i PARI 9 Us Re ANE | vil OEM PROCEEDINGS OF THE ROYAL IRISH ACADEMY VOLUME XXXIII SECTION A.—MATHEMATICAL, ASTRONOMICAL, AND PHYSICAL SCIENCE. DUBLIN: HODGES, FIGGIS, & CO., LTD. LONDON: WILLIAMS & NORGATE 1916-1917 THE ACADEMY desires it to be understood that they are not answerable for any opinion, representation of facts, or train of reasoning that may appear in any of the following Papers. The Authors of the several Essays are alone responsible for their contents. Dustin: Pemyten at THE Untyenstty Press ny Ponsonpy anv Grins. CONTENTS SECTION A.—MATHEMATICAL, ASTRONOMICAL, AND PHYSICAL SCIENCE. PAGE FirzeeraLp (Rev. R.), M.Sc., Pa.D. :— See under McCuetuanp (J. A.). Fry (M. W. J.), M.A., F.T.C.D. :— Impact in three Dimensions, : : 6 . , ¢ 0 75 Kennepy (H.), M.A., M.Sc. :— The Large Ions and Condensation-nuclei from Flames, . . 0 58 Leatuem (J. G.), M.A., D.Sc. :— On Periodic Conformal Curve-factors and Corner-factors, : , 35 McCurtuanp (J. A.), D.Sc., F.R.S., and Rev. R. Frrzaeraup, M.Sc., Pa.D. :— Photo-electric Discharge from Leaves, . 6 ; : 0 : 1 McCuetranp (J. A.), D.Sc., F.R.S., and P. J. Nouan, M.Sc. :— The Nature of the Ions produced by bubbling Air through Mercury, 24 Notan (J. J.), M.A., M.Sc. :-— The Mobilities of the Ions produced by spraying Distilled Water, . 9 Nowan (P. J.), M.A., M.Sc. :— See under McCuetuanp (J. A.). 3 woh iioiiornuns i vs co Lay me) fi Laos Phi et a x i 8 ioe mi ” ane aia Neen ; o ; : : ae iW : “i , Gy, Me hc . ll i ih Me iol tsi) a i wae airit sae te in: iy Jewett Hye): PEI Bi aN fi At int sas ie Wee a 4 suai salting yu yariqy yet 1) anil Boat he fi een 7 ao | ea Gy ‘ee ee Ae Da Hh 1 Prd ' PROCEEDINGS OF THE ROYAL TRISH ACADEMY PAPERS READ BEFORE THE ACADEMY Ie PHOTO-ELECTRIC DISCHARGE FROM LEAVES. By PROFESSOR J. A. McCLELLAND, D.Sc., F.R.S. University College, Dublin, 5) AND REV. R. FITZGERALD, M.Sc., Px.D. [Read January 2. Published May 18, 1916.] THE following paper contains an account of experiments dealing with the photo-electric properties of leaves and flowers of various kinds, and of substances extracted from the leaves. The source of ultra-violet light used was a spark between aluminium terminals placed in the secondary of an induction coil, with suitable capacity inserted in the circuit. We have not in any of the experiments described in this paper used a source of definite wave-length, but in the further examina- tion of some of the results we hope to do so. The original intention was to see if photo-electric effects could be obtained from leaves, using sunlight as the source; but many interesting points arose in preliminary experiments with the source described above, and the present paper is devoted to them. The leaves or extracts under examination were placed on an insulated support in a metal vessel, in the lid of which was a quartz window, with metal gauze underneath, through which the light passed. ‘he tray containing the leaves or solution was joined to an electrometer, and the metal vessel, the lid of which was parallel to the exposed photo-electric surface, was connected to storage cells, and kept at a high positive potential, usually 240 volts. Care was taken that the light after passing through the quartz and gauze fell only on the surface under examination, the leaves or solution covering the surface R.I.A. PROC., VOL. XXXIM., SECT. A. (1) 2 Proceedings of the Royal Irish Academy. of the vessel used to contain them so that no discharge of electrons took place from the vessel itself. The photo-electric effect is in all cases expressed in terms of the discharge from freshly cleaned copper under similar conditions. A copper plate was used as a standard, and care taken that the leaves or other surface under examination were of the same size and in same position as the standard. To obtain approximate uniformity in the light the terminals were cleaned very frequently, and all volatilized matter removed from the quartz. The electrometer used gave a deflection of 1300 scale divisions per volt, so that a small photo-electrie effect could be detected, but as the source of light was strong, and the exposed surface large, we usually had quite a large capacity joined to the electrometer. Photo-Electric Activities of various leaves and flowers. Leaves from a great many trees and shrubs, and flowers of various colours, were tested. The results may be summarized by stating that green leaves gave an effect varying from 3 per cent. to 10 per cent. of that from copper, while flowers of various colours gave smaller effects, usually less than 1 per cent. of that from copper. The following table contains a few of the leaves and flowers tested with the corresponding activities on a scale on which copper is represented by 100 :— Copper plate freshly cleaned, . : : 4 Leaves of Sycamore tree, . : : : : 10 Leaves of shrub Euonymus, : : : : 10 Leaves of Horse Chestnut, . : : ‘ : 8 Leaves of Arum lily, . : 3 é : 3 8 Leaves of Plane tree, . 7 Ivy leaves, . SS) Fresh grass, : 3 Chrysanthemum leaves,” . : ; d : 10 White Chrysanthemum flowers, . : : : 18 Pink Chrysanthemum flowers, . 5 : : 1:0 Yellow Chrysanthemum flowers, : : : 0-8 Red Chrysanthemum flowers, . : ; : 0-5 Experiments with Chlorophyll Extracts. The different behaviour of green leaves and flowers suggested that chloro- _ phyll may be the chief active agent. Hence we made experiments with chlorophy!l extracts prepared in the laboratory from various types of leaves. McCienianp ano FrrzgeraLtp— Photo-electric Discharge. 3 The chlorophyll was extracted by acetone, using in different experiments ivy leaves, leaves of nettles, and of the evergreen euonymus. We may quote the results for nettle leaves, which were similar to the others. The leaves were pounded in a mortar, put into a glass vessel, and left for some hours in contact with acetone, and the liquid then filtered off. This liquid when exposed to ultra-violet light showed practically no activity. When diluted with a large quantity of water the activity was greatly increased, the maximum effect being obtained when the acetone extract formed less than one per cent of the whole. The actual numbers for one experiment are given below :— Volume of acetone extract added to 100 c.cs. of distilled water. Activity. 2 drops, : : a3 : ; : 5 BRD 6 drops, ; : ; : : , : 6:0 20 drops = 0°5 c.c., : ; : : ‘ 9-0 65 he CO BR by Ht gt Cy Cu S&S A fresh mixture was taken at every observation, so as to eliminate fatigue effects. The activities are as before given on the scale on which a copper plate would be represented by 100. The effect was therefore a maximum when about ‘5 c.c. of this particular acetone solution was added to 100c.c. of water, and the addition of more of the solution diminished the activity. The explanation appears to be as follows:—The active substance, probably chlorophyll, is in solution in the acetone, and in this solution it is inactive. When a few drops of the acetone solution are added to the water, the substance is thrown out of solution, and shows its photo-electric effect. When more and more acetone solution is added, we reach a point when the effect of the additional acetone is to dissolve more completely the active substance, and therefore the activity decreases. This can be verified by the addition of small quantities of pure acetone at different stages. For example, in the case of one acetone solution, -2 c.c. added to 100 c.c. of water gave on an arbitrary scale an activity of 150. An additional 5 ec. reduced this to 72; but if, instead of adding the additional 5 c.c. of acetone solution, we added 5 c.c. of pure acetone, the effect was reduced to 55. These and similar observations made it clear that the photo-electric effect in these experiments depends essentially on the quantity of the active substance out of solution, as it 1s Inactive when dissolved in acetone. [1*] 4 Proecedings of the Royal Irish Academy. As mentioned above, similar results were obtained with acetone extracts from various types of leaves, and we also made similar measurements with chlorophyll purchased as pure and afterwards dissolved in acetone. Of course the percentage in the water of an acetone solution required to give the maximum effect depends on the concentration of the particular solution. Fatique Lffects of Acctone Extracts. The active substance extracted from leaves by means of acetone showed very rapid fatigue effects; and in work of the type described above, the time of exposure must be carefully attended to, if concordant results are to be obtained. The fatigue effect with these extracts is certainly far more rapid than in the case of metals. Not only is the fatigue more rapid, but there appears to be a permanent destruction of photo-electric power produced by the light to a considerable depth in the liquid. For example, in one experiment drops of acetone extract were added to water sufficient to give approximately the maximum activity. This was measured as quickly as possible and found to be 140 on our arbitrary scale. The light was kept on, and the activity fell off very rapidly. The liquid, which was 1:5 ems. deep, was then stirred up_so as to renew the surface, and the terminals and quartz plate were carefully cleaned, but it was impossible to restore the activity to more than 90. These numbers indicate a destruction of photo-electric power to a considerable depth in the liquid, and not merely a change, temporary or permanent, in the surface layer. In the case of the active substances we are dealing with, the photo-electric discharge is apparently accompanied by a permanent change in the substance itself. To test directly whether the ultra-violet rays we were using could penetrate to a considerable depth, we interposed in the path of the rays a vessel with a quartz bottom. When a layer of water 1 cm. deep was in the vessel, the rays after passing through it gave a marked photo-electric discharge from a copper plate. The transparency was, however, greatly diminished by the addition of a very few drops of the acetone extract to the water. From their known properties it is to be expected that the chlorophyll molecules would break up readily under the action of light. Experiments with water Extracts from Leaves and from Flowers of various kinds. We now decided to treat leaves directly with distilled water, and to test the water for a photo-electric effect. The first experiment tried was with McCue ianpd ANp FrrzquraLp — Photo-electric Discharge. 5 withered nettle leaves plucked some days previously. They were cut up and pressed in a mortar with cold distilled water. After filtration the liquid gave an effect corresponding to ten divisions per minute on the usual scale. After concentration to half its volume, the activity of the water was increased about 60 per cent. After standing for some time in the laboratory, the liquid was passed through a Chamberland-Pasteur filter, and found to be still active. Boiling with animal charcoal and filtering through ordinary filter paper was found to remove the activity. We tried similar experiments with leaves of euonymus, ivy, and plane tree, and we may quote the numbers for the first :— Leaves in a natural state, taken fresh from shrub, Activity = 10 Coid water extract from leaves, ‘ ; A 30 ll A freshly cleaned copper plate, : : : = 100 We were, therefore, able to get a photo-electric activity equal to 30 per cent, of that of copper by pounding up leaves in a mortar with cold distilled water. To obtain the active extract it is not necessary to break up the leaves. We immersed leaves in cold water, taking care that the surface of the leaf and the stem below the water were not broken or damaged in any way. With prolonged immersion we have obtained an activity as high as 14 on our scale. With distilled water near the boiling-point greater values are obtained, and the period of immersion may be short. Hot water poured over leaves of shrub euonymus, allowed to cool, and then tested, gave 45 per cent. of the activity of copper. The active liquid obtained in this way was found to maintain its activity for weeks when kept in a closed vessel, but the activity fell away to one-third of its value in a week when exposed to the air. We have attempted to obtain very active water solutions by concentrating weak solutions, but without any very marked success. The rate of fatigue of these water extracts was not so great as in the case of the acetone extracts diluted with water. It will be noted how much the activity of the extracts obtained from leaves by simply immersing them in water exceeded the activity of the leaves in their natural condition. Different coloured chrysanthemums, which gave a very small effect, as shown in the table above, were found to yield a water extract with an activity as high as 10 per cent. of that of copper. 6 Proceedings of the Royal Trish Academy. Effects produced by treating water Fixtracts of Leaves and some pure Organic Substances with Oxidizing Agents. Some effects that we observed led us to try the action of oxidizing agents on the photo-electric activity of water extracts from leaves, and also the action of oxidizing agents in the case of solutions of some pure organic substances. As oxidizing agents we employed lead peroxide or potassium permanganate solution. Lead peroxide, whether dry or suspended in water, is practically inactive as regards photo-electric effect, and so also is potassium permanganate. Water extracts from many types of leaves were treated with one or other of these oxidizing agents, and in nearly every case we found a large increase of photo-electric activity. Sometimes there was no increase from treatment with lead peroxide, but there was an increase when potassium permanganate was used, and in a couple of instances the increase was more marked with the lead peroxide. It is not necessary to give the numbers for all the experiments, but one or two typical cases may be quoted. A water extract from sycamore leaves had an activity of 45. When shaken up with lead peroxide and then allowed to settle, the activity was increased to 85. The rate of decay from this increased value was very slow. An extract from leaves of an ash tree had a small activity of 12, which was increased to 35 by treatment with potassium permanganate. An extract obtained from horse-chestnut leaves had an activity of 40. This was only slightly increased by potassium permanganate, but when treated with lead peroxide the activity quickly rose to 90. After standing overnight the activity had fallen very little, and when shaken up again with more lead peroxide the activity was further increased. The large increases in photo-electric activity noted above, and also results obtained when trying to isolate substances from the residues left when water extracts of leaves were distilled off, led us to try the effects of oxidizing agents on some organic substances. Hydroquinone in the solid state gave only a small effect when exposed to ultra-violet rays, but a saturated solution in cold distilled water had an activity of 40. When to the solution of hydroquinone a little paste, made by shaking up lead peroxide with water, was added, the odour of quinone was at once apparent, and the activity rose to 80. In a few hours the activity had fallen away, but was restored by adding more lead peroxide. The quinone which is produced is not specially active, and the lead peroxide is inactive. It is clear, therefore, that the increased activity is connected with the process of oxidation of the hydroquinone. McCie.nanp AND FIrzGeRALD Photo-electric Discharge. 7 Another solution of hydroquinone which we tested had an activity of 80, and this was increased when treated with lead peroxide to 115, or 15 per cent. more than a freshly cleaned copper plate. The two substances isomeric with hydroquinone, resorcine, and pyro- catechine, gave by like treatment similar results, and showed high activities. A very high photo-electric activity was obtained with pyrogallic acid. It is very soluble in water, and a solution which showed an activity of 60 had this activity increased, when shaken up with potassium permanganate, to as high a value as 400, or four times that of copper. ‘Tannin and gallic acid solutions showed little increased activity when treated with lead peroxide, but responded to potassium permanganate. As an example of a substance of a different type we tried (3-naphthol, which has two benzene rings and one hydroxyl group. Dissolved in weak alkali it showed an activity of 80, and this was increased to 170 by treatment with very dilute permanganate. Solutions in water of maltose and dextrose were tried. They gave a small activity, and it could not be increased by treatment with lead peroxide or permanganate. Methyl alcohol, ethyl alcohol, acetone, ether, aldehyde, and liquid esters were tested and found to be practically inactive. The greatest effects and the greatest increases by the use of oxidizing agents were therefore found with closed-ring compounds, the open chain compounds giving small effects. SUMMARY. 1. A table is given showing the photo-electric effects of a number of leaves and flowers. The maximum effect obtained from any type of leaf was about 10 per cent. of that from copper. 2. Acetone was used to extract chlorophyll from leaves. The acetone solution was inactive, but became active when largely diluted with water, the solid being then thrown out of solution. The photo-electric effect decays very rapidly under the action of the light. 3. It is shown that an active substance is obtained from leaves by immersing them in distilled water. When the water is near boiling point, a short time is sufficient. Large effects are obtained in this way, the water solution in some cases being half as active as copper. When comparing the effects obtained in this way with those obtained by the use of acetone, we must remember that a few drops of the acetone solutions added to a large 8 Proceedings of the Royal Irish Academy. volume of water gave a marked effect, so that the total activity obtaimed by the use of acetone greatly exceeds that of the water solutions. 4. The action of oxidizing agents greatly increases the activity of the water solutions described above, and a similar effect was obtained by adding oxidizing agents to solutions of several organic substances. Very large activities were obtained in this way. These experiments seem to afford ample proof that, in these particular cases at any rate, the emission of electrons under the action of ultra-violet light is facilitated by chemical change. It is hoped to investigate some of the points raised in this paper in a more conclusive manner, when we are iu a position to use light of definite wave- length. Dies Il. THE MOBILITIES OF IONS PRODUCED BY SPRAYING DISTILLED WATER. By J. J. NOLAN, M.A., M.Sc., University College, Dublin. Read Frpruanry 28. Published May 18, 1916. IN a previous paper* the electrification given to distilled water by breaking it up in contact with air has been investigated. It was found that the water took up a positive charge which was proportional to the area of new water- surface produced. The breaking up of the water was produced in two ways: (1) by allowing drops to splash against an air-blast, and (2) by spraying. Concordant results were obtained from these methods, the values of the charge produced per square centimetre of new surface in each case being identical. The negative charge is carried off in the air, which contains ions of both signs, negative being in excess. A knowledge of the nature of this ionisation would be of importance in any attempt to formulate a theory of the effect. The experiments described in this paper have, therefore, been undertaken. A short account of some of the results has been given already in the previous paper. A more complete account can now be given. Investigations bearing on these experiments have been made by Kahler, Aselmann,{ and Simpson.§ Kahler found that when distilled water is splashed negative ions only are produced. Aselmann also, working under similar conditions, found only negative ions. These ions had mobilities lying over a wide range. Limiting values of mobilities were obtained, the highest mobility and the lowest being respectively 4 em./sec., and 2°7 x 10+ cm/sec., in a field of 1 volt per em. Very few of the ions, however, had mobilities lower than 1-6 x 10% cm/sec. Simpson found that when distilled water is broken up in air, ions of both signs are produced, negative being in excess. No measure- ments of mobility were made. *Nolan. Proc. Roy. Soc., vol. xc, p. 531 (1914). + Kahler. Ann. der Phys., vol. xii, p. 1119 (1903). { Aselmann. Ann. der Phys., vol. xix, p. 960 (1906). § Simpson. Phil. Trans., vol, ccix, p. 379 (1909). bo R.I.A. PROC., VOL. XXXIII., SECT. A. [ 10 Proceedings of the Royal Irish Academy. There is, of course, no reason to expect that the ionisation produced by splashing and that produced by a different method of breaking up water should be identical. But apart from any apparent want of harmony between these results, it was felt that there was need for a more detailed examination of this type of ionisation. Apparatus and Method of Experiment. The arrangement of apparatus is shown in fig. 1. The spraying apparatus —a metal scent-spray fitted to a large bottle—is placed inside a cylindrical metal vessel. Connected to this vessel is the “ measuring-cylinder.” It consists of a metal cylinder with a coaxial cylindrical electrode. The latter is insulated from the cylinder, and connected to a Dolezalek electrometer. The cylinder is connected to a battery of cells. The insulation of the electrode is protected Po Electrometer B Earth Pee Fie. 1. against drops of water by a specially long guard-cylinder connected to earth. In the earlier experiments insulating material (4.4) was introduced between the measuring-cylinder and the spraying-chamber. But afterwards in working with the more mobile ions it was found convenient to establish metallic connexion, so that the two vessels were at the same potential. ‘This prevented any ions being held back by an adverse field from entering the cylinder. The measuring-cylinder was connected by tubing to a large gasometer, so that air could be drawn from the spraying-chamber (which had openings at B and () at any desired rate. In the first experiments the sprayer was driven with oxygen at a pressure of about 100 ems. of water. Most of the experiments, however, were made with air at a pressure of about 20 ems. of mercury. No difference in the nature of the ionisation was found. The results got by the two methods have not, therefore, been specially distinguished. As to the magnitude of the etfect, no proper comparison can be made, as the sprayers were different in the two cases. Notan—WMobilities of Ions Produced by Spraying. 11 The theory of the coaxial cylinder method of measuring mobilities of ions is well known. Jonised air is drawn through at a steady rate, and a current- voltage curve is plotted. Saturation for an ion of mobility w is produced at a voltage V when b Nae, (Z)e °8 Gi} © aand 6 being the radii of the inner and outer cylinders, Q the quantity of air drawn through.per second, and Z the length of the inner cylinder. If, therefore, ions of only one mobility are present, the current-voltage curve will be a straight line bending over and becoming parallel to the voltage axis at the saturation voltage. If a number of different classes of ions are present, each class having a definite mobility, the current-voltage curve is made up of vs a number of straight lines, each intersection representing a voltage at which one class of ion is saturated, and each succeeding section of the curve making a smaller angle with the voltage axis until saturation is finally reached. If the number of different classes of ions is great, the intersecting straight lines are short, and the result is, in practice, a smooth curve. But if the interval between the mobility of one class and that of the next is not too small, and if the quantities concerned are big, it is possible to use the current- voltage curve as a method of measuring the mobilities of all the ions concerned. This is the method of determining mobility that has been used throughout the work described in this paper. The ordinary procedure, therefore, was to find the rate of charging of the inner terminal for various voltages, the air being drawn steadily at a suitable rate and the pressure on the sprayer being also kept steady. In the course of the work the widest possible range of air velocities was employed, and measuring-cylinders of various dimensions were used. It is clear that the success of this method depends upon the accuracy with which the current-voltage curve can be drawn. As far as the electro- meter readings are concerned a high degree of accuracy can be reached, especially when, as in most of this work, the electrical quantities to be dealt with are quite considerable. Variations of the spraying apparatus constitute a possible source of error; but it was found that, except on very rare occasions, the sprayer was remarkably steady. But there is still the disadvantage that unless the quantity of ions of any one class is big, the bend of the curve due to the saturation of that class cannot be accurately defined, and consequently the value deduced from the mobility is subject to considerable error. This method, however, has the advantage that it can be used, without 2") 12 Proceedings of the Royal Irish Academy. any essential change of apparatus, for the measurement of widely different mobilities. General Results. An example of the sort of curve obtained is given (fig.2). This curve as drawn is made up of a number of straight lines, showing that ions of different mobilities are present. Saturation of the slowest class is effected at 266 volts, and two other classes of ions are saturated at 78 volts and 20 volts. Under the conditions of this experiment these voltages correspond to ions of mobilities “00034, -00116, and -0045 respectively. It is clear that while the bend at 20 volts and to a less degree that at 266 volts are well marked in this curve, considerable hesitation might be felt at marking a bend exactly at 78 volts. The position of this bend will toa = Satimation Velinges _- Fobiltties Zelts Fic. 2. great extent be determined by the position of one point of the curve. In fact, it might be said that a smooth curve might be drawn with almost equal exactness, showing that, instead of an abrupt step in mobilities, there was a gradual shading off from one to another, with ions of all intermediate mobilities present. With the object of eliminating this sort of uncertainty, and of obtaining as accurate values as possible for the mobilities of the different ions, the current-voltage curve was worked over in detail many times, each section being investigated under conditions specially chosen to bring out its features. Fig. 3 gives two examples of the curves drawn to determine the mobility of the slowest ion, and two referring to the ion next in order of mobility. These curves are plotted in arbitrary and different units. The values of mobility deduced from the points of bending of the curves are given. Fig. 4 gives an example of each of the next three bends of the curve and the corresponding values of the mobilities. These have been plotted to different axes and in different units for convenience, and are not to be con- Notan—WMobilities of Ions Produced by Spraying. 15 sidered with reference to the axes shown in the figure. In figures 3 and 4, therefore, we have examples of the determination of the mobilities of five different ions. oe wi | Current! ee el Mobilities. ee es 8 ese on @ ‘00088 @ -0009/ Volts Fic. 3. Current’ alts Fic. 4. It is convenient, before continuing the investigation of the current-voltage curve, to examine further the five classes of ions which we have shown to exist so far. 14 Proceedings of the Royal Irish Academy. Positive and Negative Ions. The negative current is always greater than the positive. The ratio of negative to positive when saturation is obtained is about 1:25. The five groups investigated so far are found both with positive and negative ions. The actual values of the mobilities are the same for both as far as can be determined. Effect of Time. In the case of ions derived from flames, it is found that the mobilities decrease rapidly with time. If sufficient time is allowed, the ions come to a stable mobility, having a value of about -0003 cm./sec. It is important to find if the mobilities of the ions produced by spraying depend upon the time intervening between their production and observation. TABLE Ia. Mobilities of Positive Tons. | : | | Rate of drawing air 1 2 3 4 5 | (a) 5 litresin 377 sec. | -00038 | -00107 | -0052(2) = = (2) 5 LO -00039 -00096 -0039 = == | (c) ec ie -00038 | -00088 = a ae | | @) ” 45 5 | — = | wns “0097 = (e) ”? 20 ” | a rs “0047 | = = Gy mom Ti sail i= gl ee ee Sols 045 (h) ee cSea || = | = | = 048 | TABLE Id. Mobilities of Negative Ions. Rate of drawing air 1 | 2 | 3 4 | 5 oe . : | (4) 4 litres in 150 sec. “00035 00116 | = 0045 — — (c) a HB op -00036 -00120 | -0055 | i =. (a) e 5 aes So Ze “0037 013 — (9) % 25 = — — = 1060(?) (i) am Fae Tak bs EAT (=: = |) aay No.an—WMobilities of Ions Produced by Spraying. 15 This can be investigated in two ways—(1) by varying the rate at which they are drawn from the sprayer, and (2) by keeping the rate constant, and varying the length of their path by interposing lengths of tubing. Tables are given for positive and negative ions, showing the mobilities found with different rates of drawing air. The numbers derived from experiments in which there was some doubt about the exact point of bending of the curve have been indicated by a note of interrogation. Many of the above mobilities, especially for rates 6 and ¢, are the means of a great number of determinations. These tables show, in the first place, what has been stated above—that there is no difference between positive and negative as regards mobility. They also show that the five classes of ions are present together in the air drawn from the sprayer, and that their mobilities do not change appreciably with time. Of course the agreement between the mobilities is not good in some cases; but if the positive and negative are taken together, and if allowance is made for the difficulty of obtaining good numbers for mobility, this conclusion is justified. Special care was taken in determining the mobility of the slowest ion. In this case we find a very good agreement between the mobility values for all rates. Irregularities in the mobilities of Class 2, which will be dealt with later, prevented so good an agreement in that class. The unvarying mobility of the ions is also made clear by experiments on the effect of introducing lengths of tubing between the spraying-chamber and the place of observation. The following table shows the results of one set of experiments. Hach successive observation was made after introducing about 1 metre of tubing. Under the conditions of experiment each length of tubing was equivalent to an increase of about 9 seconds in the time. TABLE II. Rate—5 litres in 75 sec. Ra Mobility of slowest Ion Proportion obrolel Tonisation due to Ya | -00038 18 per cent. 2 -00037 OB 3 00040 33 9) 4 -00039 A 5 This table shows that the mobility of the slowest ion at least is unaffected 4 16 Proceedings of the Royal Irish Academy. by time, while the percentage of the total charge carried by that ion increases rapidly. The total ionisation falls off very rapidly with time. This effect may be ascribed to recombination, which will of course be more active among the more mobile ions. An increase in the percentage of the slowest ion is therefore to be expected. Determination of the Mobilities. The results of all experiments for the determination of the mobilities of the first five classes are given in the following table. They are set down without distinction of sign. In arriving at the mean values which are given in the last line all numbers marked doubtful are excluded. Apart from these the values found for classes 1 and 2 may be regarded as having all the same TaBLe III. Mobilities of Ions of First Five Classes. 1 2 3 4 5 -00031 00087 0036" =| 013* 045* | 31 $8 | 39* -013* 048* 33 88 43* -013* 0402 33 88 45* -0102 060 2 34 88 45* 011? 37 88 48* 016? | 37 88 31 | 38 90 33 38 93 44 39 95 53 40 107 58 40 116 52? 40 120 eel 40 131 Fp. eS Unclassified Mobilities 43 | 882 -00065* Za) | -00065* 38? | 000592 44? 3 as | : - Means, | “00038 00100 | “0643 013 046 Notan—WMobilities of Ions Produced by Spraying. 17 degree of reliability. In classes J, 4,and 5, however, some of the values were obtained under conditions not specially suitable for accurate readings of these particular mobilities. The numbers obtained specially under good conditions are marked with an asterisk. The others have been disregarded in obtaining the mean values for these classes. Tt will be noted that a great number of observations of classes 1 and 2 have been made. As regards class 1 this is due to the fact that it is very difficult to decide upon the exact saturation voltage. It was considered specially important to make sure of this mobility. The same difficulty of deciding on the exact point of bending of the curve is found with class 2. But here there seems to be some real variation apart from experimental errors. The values vary betweem ‘00131 and :00088, and no less than six observations give the latter value. Then, again, disregarding one experiment about which there is some doubt, we have two experiments giving a mobility -00065. This is typical of a variation which tends to occur throughout these experi- ments—more often in the form of an increase or decrease in the quantity of some ion present, but occasionally in the occurrence of an ion of mobility which does not fall within any of the divisions ordinarily observed. The variation in quantity is very noticeable in the case of class 1. Occasionally the quantity of this ion present is so reduced that it would be difficult to assert that it was not absent. Summing up the results obtained so far, it may be said that five classes of ions have been found, each class having a distinct mobility. That these ions decrease rapidly with time, but do not alter in mobility—at least within a considerable range. The mobilities are ‘00088, -00100, 0043, ‘013, and -046 em./sec. in a field of 1 volt /em. Taking the mobility of the slowest ion as 1, these mobilities may be written 1, 2°6, 11:3, 34, and 121. Ions of Greater Mobility. In addition to the five classes dealt with above, seven other classes of ions have been found. This division into two sets of five and seven groups respectively was adopted arbitrarily at first for convenience; but, as will be shown later, there is some ground for believing that it corresponds to a definite difference in the nature of the ionisation. All twelve classes are present together; but the conditions of air-velocity and voltage which favour the observations of the more mobile groups do not yield observations of the slower ions, and vice versa. In dealing with the more mobile ions, very rapid air-blasts were generally used, and the measuring cylinder was of large diameter. Thus the time-interval between the production of the ions and their observation was R.I.A. PROC., VOL. XXXIII., SECT. A. [3] 18 Proceedings of the Royal Irish Academy. rendered very small; and, in the second place, saturation of the ions was produced at voltages high enough to enable good curves to be drawn, The negative ions are again in excess. The ratio of negative to positive is higher than in the case of the slower ions. Some of the more mobile ions are about twice as numerous as the corresponding positive ions. As far as can be seen, there is no difference between positive and negative in respect to the values observed for mobilities. Tables are given, as before, of the mobilities of positive and negative ions, classified into groups as observed with different rates of drawing air. TABLE IVa. Mobilities of Positive Ions. Rate of Drawing Air 6 7 8 9 10 11 | 12 6 litres in 39sec.) = — = “55 = a =n ie Ba Oa | = — | We - = = yes = = = 1-9 pM ” 434, | “20 — = = = ws » 32, | — = — = 1:56 = 8? ” 2 ss “14 "27 —_— —_— 1°6 = = Tasie [Vé. Mobilities of Negative Ions. Rat of Drawing Air 6 7 | 8 9 | 10 | 11 12 ) | 5 litres in 26 sec. — | — | — 1:08 | — — 6 BO my _ — | °8 1-08 ” l4 ,, = | = _ — 9) 3 — 9 es _ 22 -39 1:12 = = ES | . Ae Zee | = — 3°7 9? | | s ia ie = |) ie 1-13 = = = ; BO = — = 1-07 1:45 | 371 6-2 “5 2°25, 09 19 39 1°10 — | —_— — a Digs = yf || Ok || aoa) 1-60 3°57 6°5 + Tsihs; “13 29 =| 65) | — =| 1560 36 = | An inspection of the table of negative ions shows that the mobilities are Noian—Mobilitics of Ions Produced by Spraying. 19 not affected within the range of experiment by the rate at which the air is drawn from the sprayer. The positive ions have not been examined as fully, but there does not seem to be any ground for making distinction between the signs. At least one reliable observation has been obtained of each class of positive ions, excepting Class 12. If this ion exists with a positive charge, it must exist in very small quantities. The quantity of positive ions of Group 11 (mobility = 3) observed was very small, and was found only in one experiment. Determination of the Mobilities. The following table gives the values of mobility deduced from every experiment that was performed :— TABLE V. Mobilities of the more Mobile Groups. 6 7 8 | 9 10 11 12 12* 20* 51* 1-13* 1-4* 3-0% 54 13* 20* 51* 1-04 1-56* 3-12* 6-2 09 24* | 53% 1-08 1:56* 2°8 6:5 4 24% 53* 1-08 1-56* 2-9 al 27* 57* 1-10 1-56* 31 72 29* 39 1°10 1-60* 3-12 6-5? 19 39 Ll 1-62" 3°25 8? 22 43 112 1-62" 3:5 9? 22 45 “96? 144 3-6 mp || 6B 1-22? 1-50 3°7 28 54 1-90 3-9 55 1:90 3-0? 55 1-30 2 3-0 ? 60 1:50? = “67* ico Unclassified Mobilities, {4 + -80* - @ + 2°30 Means, 12 | 94 | 58 | 1-09 | 1-56 | 3-27 65 20 Proceedings of the Royal Irish Academy. As before, the numbers that are considered specially good are marked with an asterisk. For the classes 6, 7, 8, and 10 the means given are the means of the numbers thus distinguished. For classes 9, 11, and 12 the means given are obtained from all the results, excluding those marked doubtful. It has not been thought necessary to give examples of the curves from which these mobilities have been deduced. In general they do not differ in character from those that have been already shown. The mobilities Current Mobilis ®© 72 @ se @ Tt @® 62 © 6-5 Uelts Fig. 5. numbered 9, 11, and 12 have been difficult to establish. The quantity of each of these ions present is very small. Consequently the corresponding bends on the current-voltage curve are very faintly marked. Of the ten observations of class 9 only one can be regarded as thoroughly satisfactory. Similarly in class 11, there are only two good observations out of thirteen. Less attention, however, was given to those classes than to class 12. When indications of this class were obtained, careful experiments were made in order to verify its existence, as the mobility was considered to be remarkably high. The curves from which the values of mobility of this class are deduced are all given in fig. 5. ‘hese curves are for convenience plotted to various Notan—Mobilities of Ions Produced by Spraying. 21 axes and in various units. They all deal with negative ions. As has been already mentioned, no positive ion of this mobility (with one very doubtful exception) was observed. It is clear from the curves that the number of ions of this mobility is small in comparison with the others present. But there is hardly any room for doubt that such ions exist. The method of measuring mobilities employed during this work is not very suitable for sorting out small quantities of ions of high mobility. It is hoped that by use of a different method a better knowledge of all the more mobile groups, and especially of the last group, may- be obtained. Unelassified Mobilities. -As in the case of the slower ions, we find that certain mobilities occur which do not fit into any of the groups. Of these the most remarkable are ‘67 negative and ‘80 positive. These mobilities are deduced from very good observations, and cannot be considered to be open to doubt in any way. These observations tend to cast some doubt on our idea of coexistent groups of ious -each of definite and unchanging mobility. In dealing with the slower ions we were able to single out the ion of lowest mobility, and to show that the mobility of this class did not change with time. If Table I is referred to, and the positive and negative considered together, the reason for making this conclusion apply to all five classes will be understood. Classes 1, 2, and 3 appear on the same curve, and are found practically unchanged at rates a, 8, andec. At the last rate a new ion can be observed (no. 4). This ion, as well as no. 3, is unchanged in passing from rate ¢ to rate d. It is found again at rate f with another ion (no. 5). The latter must be a new and distinct ion. It cannot be, say, no. 1 or no. 2 in the process of growth, for we have, so to speak, kept in touch with these. As faras we can see this ion is of unvarying mobility, as there is no change in passing from rate f to rate h. If ‘lables I and IV are considered together, it will be seen that an attempt has been made to continue this process of “ keeping in touch” throughout. At each rate as many ions as possible were observed and observations were made at a great many rates. Except in this limited manner, ie., by keeping in touch with the slowest mobility through the overlapping of observations at different rates, we cannot earmark any group of ions and measure its mobility after different time-intervals. But we have at least been able to show that these groups are separate and distinct, and that the more mobile are not merely an earlier form of the slower ions. But even if each observation were isolated and 22 Proceedings of the Royal Trish Academy. separate in itself, if the process of following up the chain from the slowest to the fastest ion had not been carried out, we should come to the same conclusion. All these measurements, made under widely different conditions with regard to time, fall into a certain limited number of distinct groups. ‘There cannot, therefore, be any gradual changing of mobility from one group to another; there is no change from .013 to -0043 (Group 4 to Group 3), for if such a change occurred, intermediate values must have been observed in some of the experiments. These conclusions have special force in the slower groups (Group 2 excepted), because their strength depends upon good agreement between the values for mobility within the groups and the absence of inter- mediate values. With the exception of Group 2, these conditions are satisfied in the case of the slower ions. In the more mobile groups the gaps between the groups are small, and for some of the groups satisfactory obser- vations have not been obtained in great numbers. The occurrence, therefore, ’ of the mobilities -67, °8, 2°3, &e., tends to weaken the group system. There may be a special reason for the occurrence of these anomalous mobilities. Further investigation by a more accurate and rapider method than that employed in this work should clear up the point. The group system has been well established in the ease of the slower ions, and, in spite of the anomalies” that have been mentioned, it may be considered as hardly less firmly established in the ease of the more mobile ions also. The results obtained from these observations of the ionisation produced by spraying distilled water bear a very close relation to those obtained from experiments on the bubbling of mercury, which are given in the paper immediately following by Professor M‘Clelland and Mr. P. J. Nolan. Consideration of the results of this paper are therefore deferred. A joint discussion of the combined results is given at the end of the next paper. [SumMMARY OF RESULTS. No.an—Mobilities of Ions Produced by Spraying. 23 SUMMARY OF RESULTS. 1. When the distilled water is sprayed by air, ions of the following mobilities are obtained :— 00038, :0010, 0043, -013, -046, -12, -24, -53, 1:09, 1:56, 3:27, and 6°5 cm./sec. in a field of 1 volt /cm. 2. These ions are found carrying both positive and negative charges except the fastest (mob. = 6°5), which has not been found with a positive charge. 3. The mobility of the ions does not change with time within a certain considerable range. 4. The negative ions exceed the positive in the ratio of 5:4 for the slower ions, varying to 2: 1 for the faster groups. The writer wishes to express his indebtedness to Professor M‘Clelland, to whose suggestion this research is due, for his interest and advice. Lu 4 IE THE NATURE OF THE IONS PRODUCED BY BUBBLING AIR THROUGH MERCURY. By PROFESSOR J. A. M‘CLELLAND, D.Sc., F.R.S., AND P. J. NOLAN, MSc., University College, Dublin. Read Frepruary 28. Published May 18, 1916. THE allied subjects of the electrification produced by bubbling air through liquids and by splashing and spraying liquids have been frequently investigated in recent years, and some generalizations of importance have been obtained, but there are still much confusion and uncertainty. This paper deals with the mobility of the ions carried away by air which has bubbled through mercury, measurements being made both with air containing the normal amount of water-vapour, and also with air well dried. The paper is closely allied to the preceding paper in this volume, which deals with the mobility of ions carried away by air which has been used to spray distilled water, the object of both papers being to obtain further knowledge of the form in which the electrical charge exists. The charge of positive sign remaining on the larger drops when distilled water is sprayed has been carefully examined by J. J. Nolan,* who found that over a considerable range of size of droplets the surface density of charge was constant. Apparatus. The arrangement of apparatus shown in fig. 1 was adopted. Air was forced by a pump, driven by a motor, through a tightly packed plug of cotton wool A, about 20 ems. long, and 8 sq. ems. in cross-section. The air thus purified passed through a glass nozzle immersed in the mercury in a glass eylinder &, which was 30 cms. high, and 8 ems. in diameter. In most of the experiments the nozzle through which the air escaped was turned up as shown in the figure, and an iron plate was placed above * Electrification of Water by Splashing and Spraying. Proc. Roy. Society, A, vol. xe, 1914. ; McCuieiiand and Nouan—Tons Produced through Mercury. 25 and close to the surface of the mercury, so that the air sprayed the mercury against the iron plate. In some of the experiments the iron plate was not used, and a straight nozzle pointing downwards was employed. A part of the air which had passed through the mercury was drawn off by means of a gasometer attached to NV, through a length of tubing between C and D, and through the measuring tube J/. The surplus air escaped through the orifice O, which could be opened or closed to suit the current drawn off by the gasometer. The quantity of air drawn through the measuring tube If could be deduced from the rate of motion of the gasometer, and this rate could be varied as desired. The measuring-tube JZ consisted of a cylindrical brass tube with a metal rod 55 ems. long stretched along its axis. The tube was connected to a battery of small storage cells, and the inner terminal to an electrometer, so Electiometer (iar Me a eli We (yee Fig. 1 that the current corresponding to different voltages could be measured in the usual way. A sensitive electrometer was used giving about 1700 scale divisions per volt; but in many of the observations extra capacity was employed to reduce the motion on the scale. Any desired length of tubing could be interposed between Cand D so as to vary the time between the formation of the ions and the measurement of their mobility. The rate of flow ot the air-current through the tubing could also be varied. When dealing with the more mobile ions, a different measuring-tube was used, with greater cross-section, and with a shorter inner terminal. In some of the experiments the air was dried before being passed through the mereury. The apparatus used in drying the air consisted of two towers R.A. PROC., VOL. XXXIII., SECT. A, [4] 26 Proceedings of the Royal Irish Academy. packed with calcium chloride, and a large bulb half filled with phosphorus pentoxide. The air first passed through the calcium chloride, and then slowly over the pentoxide. The drying produced in this way was certainly not perfect, but 1t was sufficient to show the general effect produced by the partial removal of the water-vapour. Mobility of the Ions. The method used to measure the mobility consisted in plotting the current- voltage curve for a steady rate of flow of the air through the measuring-tube M. Yo obtain the saturation current corresponding to ions of mobility w a voltage V is required given by the expression Q log y a ——————— 27 Vi ? where Q is the volume of gas passing through the tube per second, } and a the radii of the tube and inner terminal, and / the length of the terminal. If there are a number of groups of ions of different mobilities present in the air, it is clear that the current-voltage curve will consist of a number of straight portions the last of which will be parallel to the axis of voltage, and the voltage for each bend of the curve will be that required -to saturate a group of ions. If the bends on the curve are not clearly defined, it will indicate that we are not dealing with distinct groups of ions, but that the range of mobilities is continuous, and the maximum current is reached for the voltage sufficient to saturate the slowest ions present. his method is, no doubt, not a very accurate one, but it is suitable for giving approximate values of the mobilities when the ions occur in a considerable number of quite distinct groups. Preliminary Observations. During the earlier observations a straight nozzle was used, the air simply bubbling through the mercury which was sprayed to some extent against the sides of the glass cylinder. It was observed that the mobilities varied with the speed of the air-current and with the length of tubing inserted between the mercury vessel and the measuring-tube MZ. The longer the time that elapsed between the bubbling through the mercury and the measurements, the slower were the mobilities. When the length of tubing was sufficiently great, the mobilities reached steady values. The current was fully saturated at a voltage corresponding to a mobility of -00034 ems. per second for a gradient of one volt per cm. McCuietnanp anv Notan—Zons Produced through Mercury. 27 This is the mobility of the large ions found in the atmosphere and of the large ions in flame-gas which has been allowed to cool. The slowest ion in air which has bubbled through mercury reaches this same mobility after some time, and apparently does not suffer any further decrease of mobility. The curves for positive and negative ions were similar, and the slowest ion is the same in the two cases. ~ In addition to the ions of mobility -00054, it was evident from the bends in the curve that other distinct groups were present. The amount of the ionization was, however, small, especially when time was given for steady mobilities to be reached, and we endeavoured to increase the ionization before making a detailed examination of the curves. The amount of the ionization depended on the manner in which the mercury was thrown about by the air-blast; but when the disturbance is rather violent, it is difficult to obtain steadiness. After trying various types of nozzles we used a turned-up nozzle a very short depth below the surface of the mercury, and placed an iron plate just above the surface, so that the mercury was blown against the plate. ‘his arrangement we found gave a very large supply of positive ions, but the number of negative ions was practi- cally unchanged by the presence of the iron plate. We afterwards used this arrangement, and most of the measurements were made with positive ions, as their greater number made the observations easier and more accurate. The mercury was purified well when beginning the observations, but the constant bubbling of air through it brought out further traces of impurities, and these were removed at later stages. It was noticed that the ionization decreased as the mercury became purer. We decided to determine carefully the mobilities of the different groups of ions in the following cases: —(1) with undried air and giving sufficient time to allow the mobilities to reach their minimum values; (2) with dried air and allowing the same time to elapse between the production of the ions and their measurements; (3) and (4) with undried and dried air respectively, the measurements of mobility being carried out as soon after the bubbling as convenient. Ondried Air. Constant Mobilities. The curves on figs. 2 and 3 are examples of those we plotted when the time interval between the mercury vessel and the measuring-tube was sufficient to allow the mobilities to reach steady values. In both figures the curves are plotted in two parts, the smaller numbers on the axis of voltage referring to the lower curve. The rate of air-current through the [4*] 28 Proceedings of the Royal Irish Academy. measuring-tube is greater for fig. 2 than for fig. 5. The calculated mobilities corresponding to the well-marked bends are given. The agreement is fairly good, considering the nature of the observations. The time which has elapsed from the production of the ions to the measurement of their mobilities is about 60 seconds in the case of fig. 5 and 30 seconds for fig. 2. In neither of these curves is the bend shown corresponding to the mobility -00034 ems. per second. This was carefully determined by observations with slower air- currents. @ 00s @-oo04 @o1#0 ° fog 200 Joo 400 a 70 Volts 60 0 Fig. 3 The curves do not pass through the origin, which indicates that there is at least one further group of ions of mobility greater than -0015 cms. per second. In later work it will be found that five sets of ions were shown to McCurtianp and Notan—LTons Produced through Mercury. 29 exist. The means of a number of observations of the type illustrated by these curves gave the following mobilities :— 00034, ‘0013, ‘004, 014. The ratios of these mobilities are .— 1 3:8 11°8 4]. As stated above, all the observations in this section were with positive ions, except in the case of the slowest group, in which ions of both signs were measured, and found to have the same mobility. Dried Air. The air was now passed through drying tubes before reaching the mercury, and the same time interval allowed between the bubbling and the measurements as in the above experiments with undried air. One result of drying the air was to reduce the amount of ionization to about one-sixth of its value with undried air. The mobilities were greater than with undried air, although the time interval was equally long. Measurements of the mobilities made at different times did not now show such a satisfactory agreement, possibly because the degree of dryness on different occasions was not the same. One set of experiments gave the following mobilities :— 00056, 0021, 0068, ‘024 cms. per sec. The ratios of these mobilities are :— iL, 38, 12:1, 42°8. The observations to which these numbers refer were made with positive ions only. The question arises as to whether these ions had reached constant mobilities, or whether they were still decreasing in mobility towards the values found in the preceding section. Our observations show that after the interval of time allowed the mobilities, if not constant, were changing very slowly. Dried Air. Mobilities measured before stable condition was reached. Continuing to use dried air, we now measured the mobilities as soon as convenient after the air bubbled through the mercury. As higher mobilities were now found, we used a measuring-tube with a shorter inner terminal. The dimensions of the apparatus and air-blast were such that about 10 seconds interval occurred before the measurements were made. A careful examination of the air showed that the fastest moving ion had a mobility of °32 ems. per second. 30 Proceedings of the Royal Irish Academy. With this measuring-tube three distinct bends on the current voltage curve were determined, corresponding to mobilities ‘32, 092, and 043 cms. per second. With the larger measuring-tube previously used the slower ions were examined, and two groups were found with mobilities 0064 and :0022 ems. per second. With a third measuring-tube of intermediate size the portion of the curve corresponding to mobilities between 043 and -0064 was examined, but found to be free from bends, showing that there were no groups of ions with mobilities between these values. We have, therefore, five distinct groups of ions with mobilities :— 0022, ‘0064, 043, 092, and -32. The ratios of these mobilities are :— il 2:9, 1G), 418, and 146. These numbers are the means of several observations, which showed fairly good agreement. Corrvent im Uoll's Fic. 4. As the ionization was greater under the present conditions, we were able to make observations on both positive and negative ions, and in the case of all the five groups the ions of different signs were found to have equal mobilities. The mobilities given above are greater than the corresponding numbers when a longer time interval was allowed. Probably even larger numbers would have been found if the measurements had been made with a still shorter time interval. McCietianp and Notan—Tons Produced through Mercury. 31 It will be observed that the ratios of the mobilities are not very different from the ratios in the preceding sections. Further, it will be observed that five groups are present, and in the preceding sections, while we have measure- ments for only four groups, there was evidence of the presence of a fifth group. To give an idea of the distinct character of these groups, we show on fig. 4 portions of some of the curves from which the mobilities are calculated. The curves are plotted on arbitrary scales, and without reference to the origin. They are given to show that the curves on both sides of each bend are straight lines. The upper curve shows the bend corresponding to the slowest lon present. Undried Air. The mobilities were also measured about ten seconds after the production of the ions, using undried air. In this case also five groups were found. The positive and negative ions in each group had the same mobility, and the mean values obtained were : — 0013, 0045, 02, 048, “20 The ratios are :— il 34, 15-4, 37 154 The mobilities are less than for dry air with the same time interval, and the amount of ionization was greater. The positive ions were more numerous than the negative, the excess occurring chiefly in the groups with mobilities -02 and ‘048. Some observa- tions were again made with the iron plate removed, and using a straight nozzle, the other conditions being as in this section. The mobilities were practically the same as when the mercury sprayed against the iron, but the number of positive ions was greatly decreased. SUMMARY. The results in the preceding sections are collected in the following table: Mobilities of Groups. Me II. JOU IV. W Undried air: Long time interval, . -00054 ‘0013 “004 “014 = Ratios, . : : 1 38 11°8 41 — Dried air : Long time interval,. -00056 0021 0068 024 = Ratios, . : : 1 38 12:1 42°8 _— 52 Proceedings of the Royal Irish Academy. Undried air: Short time interval, “0013 “0045 02 “048 20 Ratios, . : : 1 3-4 15-4 a7 154 Dried air: Short time interval, “0022 0064 043 “092 3) Ratios, . : : iL 2°9 19-5 418 145 As it happens, the shorter time interval employed in obtaining the numbers in this table was such that the mobility of the slowest group was the same as that of the second group when stable, and other coincidences of the same nature occur in the table. It should be remembered, however, that with the short time interval the mobilities are not constant, but depend on the time. We should not, therefore, specially identify a stable ion in Group II with a varying ion in Group I which happens to have the same mobility. The ratios of the mobilities in the different groups are approximately the same, whatever the time interval, and whether the air is dried or undried. No doubt, very considerable variations occur in these ratios, but, considering the difficulty of deciding on the mobility in many cases,-the approximate constancy is striking. JOINT DiIscussION OF RESULTS WITH THE AUTHOR OF THE PRECEDING PAPER ON THE Mopitity oF IONS PRODUCED BY SPRAYING DISTILLED WATER. The experimental results in this paper are so closely related to those in the preceding paper that it is convenient to discuss them together. When distilled water is sprayed, the air carries away ions of many distinct groups, the five slowest of which correspond closely with the stable groups found in the experiments with mercury in undried air. The more mobile ions found in the experiments with water are either entirely absent from or present only in very small numbers in the mercury experiments. We shall first discuss the probable nature of these five groups which are common to the two sets of experiments. The differences in the behaviour in the two cases are apparently due to the fact that those produced from mercury take some time to add on water- vapour, and thus arrive at a stable condition, while the others produced from water do not show any change of mobility with time. When the air is “dried” in the mercury experiments, all the water-vapour is not removed, but the approach to stable conditions is retarded by the diminution of the vapour present. Various hypotheses might be suggested to account for the McCretianp and No~an—Jons Produced through Mercury. 38 mobilities of these different groups of ions and for the constancy of the ratios of their mobilities during the process of growth. 1. We might assume that all the ions are the same size at any instant, and will, therefore, take on water-vapour at the same rate, and in this way we can get an explanation of the constancy of the ratios. We would then, however, have to explain the different mobilities by different charges, and it is difficult to see how different charges could possibly give us the steps in mobility we have observed. It is extremely improbable that the correct explanation can be found on any assumption of variable charges on similar nuclei. 2. We might assume that the ions consist of water-globules of different sizes haying the same charge, or we need make no assumption regarding the charge if the mobility of such ions is approximately independent of the charge. The fact that both the mobility and the rate of taking on water would depend on the size of the globule might result in the ratios of the mobilities being approximately constant. The five separate ions in their final steady state would from this point of view be five globules of different sizes, each possessing some degree of stability. The objection to this view is that we have ions proceeding past certain apparently stable sizes until each arrives at its characteristic size. The difficulties in the way of this hypothesis are very great. 3. We can modify this assumption (2) in a way that removes the serious difficulties in accepting it. Let us assume that there is one stable size of water-globule, and that the five different ions consist of groupings of different numbers of these globules. Before the steady state is reached each globule is taking on water, and, therefore, the grouping which constitutes an ion is growing at a rate depending on the number of globules it contains. The constancy of the ratios of the mobilities is at once explained on this theory. As an example of how such groups may be built up, we may start with a single globule carrying a unit charge and having a certain mobility. The next ion may contain a number of these globules, say three, two positive and one negative, or two negative and one positive. Such an ion might have approxi- mately one-third the mobility of the single globule. Similarly some grouping of these ions might form a still more complex and more slowly moving ion, and so on. It may be noticed that the average of the observed mobility ratios is about 3°4. It is possible that on some such lines as are here indicated an explanation of the different groups of ions may be found. According to this view ions of each class combine to form the next slower class, and if sufficient time is given we should have an excess of the slowest and most complex ion. Certainly R.I.A. PROC., VOL. XXXIII., SECT. A. [5] 34 Proceedings of the Royal Irish Academy. the percentage present of the slowest ion increases with the time ; but whether this is due partly to the reason here suggested, or whether it can be sufticiently explained by the more rapid loss of the more mobile ions, we are not yet in a position to say. Further work is also required in the direction of measuring the mobilities at shorter intervals after the spraying and the bubbling through mercury. In this discussion we have dealt only with the five groups of ions which oceur both in the experiments with water and with mercury, the slowest of these ions being the well-known large ion of Langevin. The other groups of ions found in the spraying of distilled water contain the ordinary small ion and at least two classes of even greater mobility. These ions are now being further investigated by a different and more convenient method, and we must postpone further discussion for the present. It would appear that a fuller knowledge of the different groups of the more mobile ions that occur in these experiments should help to elucidate the nature of the ordinary small ion, while a study of the slower groups can hardly fail to add to our knowledge of the Langevin ion. (1 35 1] VE ON PERIODIC CONFORMAL CURVE-FACTORS AND CORNER-FACTORS. By J. G. LEATHEM, M.A., D.Sc. Read Fepruary 14. Published Aveusr 11, 1916. 1. Introduction.—In a previous paper* the writer has defined conformal curve-factors, and exemplified their use in the conformal representation of simply connected two-dimensional regions of assigned type, say in the plane of a complex variable z = 2 + zy, upon the principal half-plane of a variable w= + Ww. If there is justification for the hope that the method of curve-factors constitutes a more systematic and comprehensive mode of approach to those classes of physical problems which can be formulated in terms of conformal transformation than any previously recognised method, it is worth while to consider how it may be extended to the conformal representation upon the principal half-plane of w of such a doubly connected region in the z plane as is unbounded externally but is bounded internally by a single closed curve, not necessarily free from corners. Such a representation would find illustra- tion in the circulatory irrotational motion of liquid round a fixed internal boundary, the velocity being the downward gradient of @, or in the electro- static field round a charged cylindrical conductor, the electrostatic potential being — 1p. If the hydrodynamical circulation round the cylinder, or alternatively its electric charge per unit length, is to be definite, the inner boundary of the field of flow or induction will correspond, not to the whole real axis in the plane of w, but to a definite length > upon it, which may be called the “ linear period.” The complete half-plane of won the positive side of the real axis corresponds to the doubly connected region in the z plane, repeated again and again, and z is a periodic function of w having the real wave-length or linear period X. Also dz/dw is a periodic function of w. The transformation, in its differential form, is therefore of the type *Some Applications of conformal transformation to problems in Hydrodynamics. Roy. Soc. Phil. Trans. A., vol. cexv, 1915, p. 439. R.1,A. PROC., VOL. XXXII, SECT. A. [6] 36 Proceedings of the Royal Irish Academy. dz = ( (w)dw, where ( (2) is a curve-factor periodic in ¢% with linear period X. And, as a tangent to the internal boundary makes one complete revolution as the point of contact goes once round the boundary, the angular sub-range of G corresponding to the linear sub-range X is 27; this fact may be conveniently expressed by saying that the “angular period” of @ is 27. It is proposed to look for types of periodic curve-factors which can be employed to give conformal representation of doubly connected regions whose only boundary is internal, and it will be seen that such curve-factors may be used to construct not only differential transformation formulae but also formulae in which z is expressed explicitly in terms of w. 2. Circular and elliptic curve-factors and derived types——W hen the boundary is a circle of radius a, the origin of z may be taken at the centre, and the field of flow or induction is determined by the relation w = (tA/27) log (z/a), (1) where the sign is so chosen that @ increases as the circle is described in the clockwise sense, which implies keeping the relevant region on the left. The relation is equivalent to the differential formula dz = — (27ria/X) exp (— 27riw/d) dw, (2) and this gives the periodic curve-factor 6 = exp (— 2rriw/d). (3) When the boundary is an ellipse of semi-axes ccosh a, csinha, with centre at the origin of z and major axis along the real axis, the field is determined by the relation 2 = ccosh [— (277/A) w + a]. (4) The corresponding differential relation, dz = — (2rie/X) sinh [- (2772/A) w + al, (5) gives the curve-factor Gs, = sinh [= (2m7/A) w + aj. (6) G;, and (¢,, have no zeroes or infinities for definite positive values of yp. Their only infinity in the relevant region is for yr—> + oo, and that of course corresponds to the external boundlessness of the relevant region in the z plane. As periodicity with a linear period which is a submultiple of A implies periodicity with linear period \, the substitution in @;, and G3 of X/n for X, where 7 is any integer, will give periodic curve-factors. In the case of @;,; the substitution leads simply to the nth power of @;,, and so does not give a new type. But : Gy = sinh [— (27ni/A) w + a] (7) is not a mere power of (;, and is therefore a new type. Its angular period Learurm—Periodie Conformal Curve- Factors and Corner-Factors. 37 corresponding to the linear sub-range A is 2n7, so that if an angular period of 2a were desired (1/",, might be employed. One of the important characteristics of the special kind of conformal representation now under consideration, namely the periodicity of dz/dw with linear period X and angular period 27, can be secured by making dz/dw_ pro- portional to G'/"5 or to a product of powers of two or all of the types Gx, Css; Gs, provided the angular period of the combination is 27. For example, 6? 5; 6% or G55 6%y is, to this extent, applicable provided p + ng = 1. The sum of two curve-factors is sometimes, but not always, a curve-factor. For example, the function be = exp (- 2ariw/d) + k exp (— 2aniw/d), (8) can only vanish when the moduli of the two complex terms are equal, so that exp (2a/A) = | & | exp (Qnmp/A); for positive W this implies 1=|k| exp {2(1-1)mp/A}, which is impossible if | &|>1. Hence, when this inequality is satisfied, Go is a curve-factor. As regards angular period, if one represents the two terms as vectors which are to be added by the triangle law, one readily sees that the angular period of the sum is the same as that of the term which has always the greater modulus. So the angular period of Gy is 2nz. Similarly it can be seen that Gm = @ exp (= 2antw/d) + 6 exp (- 2arntw/A) + ¢ exp (— 2ansiw/A) (9) is a curve-factor provided that, for all positive values of y, | a | exp (2amW/A) > | 0 | exp (2rnW/d) + | ¢ | exp (27nf/A), (10) as, for example, when 7, > %, ™ >, and |a| > |6|+ |c|; the angular period is 2777. 3. Transformations which are not in differential form.—There is another way of employing periodic curve-factors for obtaining conformal representa- tion of regions of the kind under discussion. If such a representation be specified by a formula z=f(w), the origin of z being supposed inside the closed boundary, the function / has to satisfy three requirements :—(1) / must have no zeroes or infinities for positive values of yy, save an infinity for y->+o. (2) f must be periodic in w with real linear period X. (3) The periodicity of f must be such that, when a point traverses a length A of the real axis in the w plane, the corresponding point in the z plane describes a closed path which encircles the origin once and only once. Now any periodic curve-factor 6 (w), whose linear and angular periods are A and 27 respec- tively, satisfies all these requirements. Hence z= ¢(w) specifies a conformal representation of the kind of region desired. [6*] 38 Proceedings of the Royal Irish Academy. Similar use might be made of a periodic curve-factor of zero angular period, the origin in the plane of z being outside the boundary curve. The relation between the boundaries corresponding to dz = ((w)dw and to z= (w) is analogous to the relation between an orbit and its hodo- graph. The determination of the form of the boundary is usually easier when the transformation is of the latter type, as might be exemplified by taking @ to be O'/" or ('l",, But the advantage, at the present stage of the discussion, of transformations of the type z= (w) is that they give representations possessing not only one characteristic but all the characteristics required by the specification of article 1. 4. Condition for the periodicity of s.—In the previous article it has been seen that the conditions which must be satisfied by @ include all the con- ditions which must be satisfied by 7, The converse theorem, however, is not true, and the difference is important. One characteristic feature of the problem under consideration has been formulated early in article 2. Another characteristic feature is that the boundary (corresponding to y = 0) in the z plane is a closed curve, and that the curves which correspond to positive constant values of y are also closed curves. In other words z is periodic in w with linear period X. lf the differential relation dz = G’dw lead, on integration, to the relation z=F(w), so that F=f (dw, it is necessary that both @ and F be periodic. But the mere periodicity of Gis not a guarantee of the periodicity of #’; for if a constant g (possibly complex) be added to @, the periodicity of Gis not impaired, while a non-periodic part gw is added to #. Thus, in the absence of precaution to the contrary, there is always a chance of a periodic @ leading to a non-periodie #. In the geometrical interpretation this would mean that the curve in the z plane corresponding to ~ = 0, instead of being a single closed loop, would be an infinitely extended periodic curve, necessarily with nodes and loops when the angular period is 27, of the general character, for example, of a nodal trochoid. It may obviously be said of such a curve that dz is periodic, but z is not. If C be the mean value of G calculated for a fixed value of ¥, zero or positive, and for a range of values of @ of extent , then for the corresponding range {| Gdw-=XC, and the value of z does not repeat itself unless U = 0. Hence there must be added to the conditions which @ has to satisfy the requirement that its mean value as here defined must be zero. It can be seen that, provided @ is periodic, the mean value which has been defined is independent of the particular positive constant value assigned to yp. For, as @ has by hypothesis no singularities in the half-plane of y positive, LrearHem—Pertodic Conformal Curve- Factors and Corner-Factors. 39 { @dw round any contour in that region is zero. Let the contour be a rectangle whose corners are dot th, dt+rAt+ Mh, GtA+ Me, gH + Wr through its periodicity the subject of integration has equal values at corre- sponding points on the sides lying in the lines ¢=9,, @=9 +A, and so the corresponding contributions to the contour integral vanish; hence the parts of the integral corresponding to the sides in the lines ~=y,, ~=y.2, add up to zero. Thus { Gdw has the same value when taken from @=¢, to d=¢,+A with constant ~, whether ~=y, or ~=y». It follows that the mean value of 6” is the same for both ranges. This suggests a method of formulating the condition for a closed curve, that is the condition for periodicity of z or { Gdw, which is useful in many eases. It consists in getting the mean value of (for ~ > + ~, and equating it to zero. Suppose that, for great positive values of ~, G can be expressed as a series of descending integral powers of exp (— 2miw/A), say s= 20 G = exp(—2nniw/d) > c, exp (Qasiw/A), (11) s=0 and that the series is integrable for ~—- +0. Then, 2 being an integer, it is to be observed that every term is periodic and has the mean value zero except that corresponding to s=n, which is a constant. Thus the mean value of @ is c,, and the condition for periodicity of { Gdw is c, = 9. By way of illustration, let the test be applied to 61/",. This can be put in the form : kt)” exp (= 2riw/d) [1 + k7 exp {2a (nm — 1) w/d} J", (12) and the binomial expansion is valid for great positive values of y. If n = 2, there is a constant term in the expansion, but for any other integral value of n there is none. Thus the transformation dz= 64/"»dw gives a closed curve for any integral value of » exeept n = 2. Similarly n , ,, 1jn G15 = (3). exp (- a + ‘| E ~exp (= = 2a) ae (3) cA and the binomial expansion is valid for great positive values of ¥. There is no constant term for any integral value of n, and so the transformation dz = 6'/";,dw_ gives a closed curve. 5. Relation between angular period and exponential order at infinity.— When a periodic curve-factor can be expanded, for ~ great and positive, in the form indicated in formula 11, it may be said to have a definite “expo- nential order at infinity,” namely 2n7/A, this being the coefficient of y in the 4() Proceedings of the Royal Irish Academy. exponential term of highest order. The exponential order at infinity is closely related to the angular period. There being no definite infinities or zeroes of (in the relevant region, the integral {d¢’/¢ taken round any contour is zero. Let the contour be the rectangle formed by the lines ~=0, P=t, $=, $= +A; since Gis periodic, the subject of integration is equal at corresponding points of the lines ¢=% and ¢=9,+X, and so the integrals along these sides of the rectangle cancel one another. The integral along the length X of the line ~=0 equals —7 times the corresponding angular range, and the integral along the length X of the line Y =¢ equals + 7 times the angular range for this line, that is the angle between the tangents at the extremities of the corresponding curve in the z plane. Thus it appears that the angular range is the same for all lines of length X parallel to the line ~ = 0, being in fact equal to the angular period of @. If ¢ be made indefinitely great, the limit value of the integral depends on the term of highest exponential order in the formula 11. If be put equal to Aexp(—iNw), then dé/G =-iNdw, and the integral from ¢, + iz to p.+AX+uz is -iNX. So if / have its exponential order at infinity V and its angular period Q, WA=Q. When Q=27, N= 2z7/d, and therefore the x of formula (11) must be unity. 6. A more comprehensive formula for periodic curve-factors.—As the number of types of periodic curve-factor as yet obtained is small, it is desirable to seek some wider formula which may be used for the extension of the category of known types. Consider a semi-infinite strip of width A in the w plane, say the strip between the lines ¢=—3X, ~=4)A, on the positive side of the axis of #. Se If (is (i) periodic of linear period X, (ii) free from definite zeroes and infinities in the half-plane of w, then it i equally general to describe ( as (i) periodic of linear W=-2X W=inr period X, (ii) free from definite zeroes and infinities in the = chi strip. The strip in the w plane can be represented conformally upon the half- plane of a new variable @ by the transformation 6 = csin(zv/d), (14) where ¢ is a real constant. Now @; having no definite zeroes or infinities in the strip in the w plane, must, when expressed as a function of 6, be free from definite zeroes and infinities in the positive half-plane of @. And as ( is a curve-factor whose range of curvilinearity covers A on the real axis in the LearnemM— Periodic Conformal Curve-Factors and Corner-Fuctors. 41 w plane, and has no branch-points on the lines ¢=+ 3A (save possibly at the corners of the strip), so-@ is a curve-factor in 4 whose range of curvilinearity does not extend outside the range from — ¢ to + ¢. If the range of curvilinearity covers only part of the range from —¢ to +c, the curve in the z plane corresponding, in the transformation dz= Cdw, to y= 0 will have one or more straight portions, without loss of smoothness at points between the points w=+ 4X where curved and straight portions meet. In such cases G is not a simple curve-factor. It is to be noted that @ is periodic in w, of linear period 2A. Therefore any function of 6, defined so as to be single-valued over any region of the @ plane, is, when expressed as a function of w, periodic of linear period 2 within that region. So, in particular, if the plane of @ be cut along the real axis from §=—c to @=c, any curve-factor in 9 whose branchings are all in this cut, and which is single-valued in the cut plane, is periodic in w with linear period 2X. But what is required of @ is periodicity of linear period X, so that not all curve-factors in @ satisfy the requirement. Thus the attempt to generalize has led to the following verbal formula for a periodic curve-factor:—Any curve-factor in the variable @ = ¢ sin (ww/X) which is. periodic in w with linear period A, and has the linear range — ¢ to ¢ or any range within that range. As regards the required periodicity, it is to be noticed that the addition of X to w changes 9 into — 6, so that 6’ must be a function of 4 whose value is unaltered by change of the sign of 0. But, seeing that the variables dealt with are complex, and that there may be branch-points or a continuous distribution of branching along the range from 6=-—c to #=c, the effect upon @ of a change in the sign of @ cannot be estimated by a mere glance at the functional formula, but must be studied more closely. If the positive half-plane of w be divided up into a continuous series of semi-infinite strips of breadth A, one of which is the strip from ¢=- $A to @=44A, the transformation (14) represents only alternate strips of the series upon the positive half-plane of 6. The other strips, including those immediately adjoining the above specified strip, are conformally represented on that half-plane of @ for which the imaginary part is negative. The addition of A to w involves a passage from a point in one strip to the corresponding point in the next strip; but this passage must be along a path which does not cross the axis ~ = 0, and therefore does cross the boundary between the strips. The corresponding change is from a value @ to a value — @; but in the @ plane the passage is not along any arbitrarily selected path, it must be along a path which does not cut the part of the real axis between 6=—c and 6 =c, but crosses the real axis somewhere outside that range. 42 Proceedings of the Royal Irish Academy. A type of path from @ to — @ is represented in figure 2, wherein 0 is the origin of 6, and A, 4’ are points in the axis equidistant from 0. The path from A to A’ is in the axis, except for semi-circular detours round branch- points between — ¢ and c¢, and a circular detour round ¢, which may be a branch-point. Fietre 2. The function £(@) may have one or more groups of branch-points, such that within each group the powers of the branchings are additive.* Attention being directed to the branchings of such a group, it is known that each semi- circle of detour round a point where there is branching of power a introduces a factor exp (-iza) into the corresponding part of the function. Now A may be taken anywhere from 0 to ¢, and if it be possible by moving A to introduce or remove a semi-circular detour in the part of the path from A to ¢ without making simultaneously a corresponding change in the semi-circular detours in the part of the path from 0 to A’, then it is impossible for the effect of the traversing of the path from 4 to A’ to be independent of the position of A, as it must be if ¢(A)=€(A). Hence it is necessary that, within each additive group, the distribution of branch-points along the range from —c to ¢ be symmetrical with respect to 0. This symmetry once recognized, it is seen that to each branching of power a in the range from 0 to ¢ there correspond two semi-circular detours in the path, either two at the same point, if between A and c¢, or two at points symmetrically situated with respect to 0, if between A’ and 4. The only exception is the branch-point (if any) of power a, at 0, for which there is only one semi-circle. As the function is to have the same value for — 0 as for 6, the cumulative effect of all the semi-circular detours corresponding to the branchings of an additive group must be the restoration of the original * This may be explained by an example. In the case of the function 62+ 1+ (@ — a2)? + (0 — Bj: (a — )3 (0 — 3, the powers of the branchings at 6==+« may be regarded as additive, since they affect the same term of the function; and the branchingsat @=6, @=¢, @=4@ forma group whose powers are likewise additive. Lea tHrm— Periodic Conformal Curve-Factors and Oorner-Factors. 43 value; hence, as each contributes a factor exp (- ima), it is necessary that eXp {— Um (a) + 22a)} =1, or a, + 23a = an even integer. In this statement, on account of the exceptional circumstances at O, an integral odd power of # must be regarded as a branching; thus, to take the simplest possible example, the function 2 24 (@ - 40) changes sign with 0, but the function does not. It is of interest to see how some of the already known types fit into the formula. It is easy to see that 7 ; (? 6/0? \t 2 ono Dia) aha DE OLS r 6x = exp (— 2mw/X) = 1 Z (3 Ly : (15) and that Bp = - +¢ +0 (0° = cy? (16) is a curve-factor in @ of linear range — ¢ to ¢ and angular range 27. Similarly it is seen that G3 = sinh {— (Qiw/A) + a} 2 2 i : (1 9 3) sinh a - 29(% - 1) cosh a, (17) where a Cx, = (0° — 4c?) sinha + 0 (6? — 2)? cosha (18) is a curve-factor in @ of lineay range - ¢ to ¢ and angular range 27. Another suitable curve-factor in @ is C= BP - 0) + (6? — a2)? (8° - 0/3, (19) where a 0 is readily verified. 8. Conformal representation of the space outside a triangle or closed polygon.— Let the angles of a triangle in the z plane be A, B, C, and let the values a, 3, y be assigned to # at the corners. By the previous article if appears that the conformal transformation of the region outside the triangle is deter- mined by the formula A B G U2 ale Gn, (grace ia Nae rete TE Nr) ay ee K sins (w- a), ion x2 - 2), ee x (u =a) , (22) where X is a constant,and A>y-a>f—-a>0. The obvious periodicity of the expression on the right-hand side of the formula is for a linear period 2 ; but when it is noticed (i) that the modulus has a period A, (ii) that the decrease of vector angle as w passes through real values from say wp to w, + A is T—-A +7 —- B+7—-C =2n, it becomes clear that the expression is periodic with linear period X. In general, though the transformation (22) gives a periodic dz/dw, there is no reason why it should give a periodic z. Usually the value of z for w=a+A will be different from that for w = a, and the boundary will be a continuous recurring rectilineal pattern having the kind of periodicity that would be got by printing from a rolling cylinder on a long straight ribbon, namely a space periodicity with respect to z. But, if a particular relation subsists between the parameters, z is a periodic function of w, and the boundary is a triangle. Though the right-hand side of formula (22) is not a proper curve-factor, the method of article 4 for obtaining the condition that z be periodic is applicable to it. The condition is therefore the vanishing of the absolute term in that expansion of dz/dw in ascending powers of exp (2miw/A) which is valid for great positive values of yp. On putting Tw r sin ie (a-w) f — exp an (w - ai Nn WV? and employing the binomial theorem, it is seen that the expansion in (w - a) = $0 exp question is _ (tm AY) 2imw\ ( =) ( 2ima ] Sit ie aie es a ee ss LK exp | y Za(1 =~) exe X } (i — } exp X Jee} 0 (23) where the terms that should follow the final plus sign are positive powers of exp (2miw/d). Coy 46 Proceedings of the Royal Irish Academy. Thus the condition for periodic z, that is, for a closed triangle, is if A {2am = Nex = 0) 24 za a) A ) This is, of course, equivalent to the two conditions 3(1- *) cos (=F*) = 0 =(1 -=) sin (772) = 0, (25) i and these are equivalent to the statement that a triangle can be drawn such that the directions of its sides make the angles 27ra/X, 27 /A, 2Qary/X, with a fixed direction, and the lengths of the sides are proportional to 1 - (A/a), 1-(B/z), 1-(C/r). All the above argument applies equally to a polygon whose internal angles are A, B, C, D, &c., provided each of the summations be understood to include as many terms as there are corners. 9. Focal periodic curve-factors.—The transformation of the previous article, giving the field outside a polygonal boundary, may be interpreted as giving the field round a different boundary got by assigning to y a constant positive value. Such new boundary would be smooth. A new yf, chosen to vanish on the new boundary, is introduced by substituting ~ + « for y, or w+ for w, in formula (22), it being understood that « is positive. Thus there is obtained what may be called a focal periodic curve-factor, A = I [sin {7 (w + %& — a)/A} 5, (26) wherein amet and =(1 -=)exp(-=2")- 0. (27) \ 7 The corners of the polygon, now outside the relevant region, may be called foci of /.;, and any number of foci may be introduced. When there are two foci, d = 0, Bb = 0, and the condition for periodicity of z is exp (— 2i7ra/A) + exp (— 2777 8/A) = 0 which is equivalent to @=a+4X. So the bifocal curve-factor is seen to be practically equivalent to @;:. In /,, it is understood that « is the same in all the factors of the product, but if, instead, the constant typified by « be different in the different factors, and denoted by a’, 8’, &c., there results the more general function f A Ges = I [sin [aw (w - a + ta’)/X} 5, (28) subject to the conditions oat S)) So) 3 (1 -2)exp }-5 <= (a- ie’)} = 0. (29) Learnem—Periodie Conformal Curve-Factors and Corner-Factors. 47 This is as free from zeros and infinities in the relevant region as is @,, and is clearly periodic of-linear period X. It is therefore a periodic curve-factor, and has the angular period 27, A curved boundary with one or more corners is given by the special form assumed by Gs when one or more of the constants a’, B’,... is taken to be Zero. 10. Periodic curve-factors regarded as limits of products of perrodic corner- factors.—The field outside a closed polygonal boundary being obtained by the method of article 8, it is possible to increase the number of sides of the polygon without limit in such manner that the polygon tends to a limit form which is a smooth closed curve. The corresponding limit of the product of corner-factors which takes the place of the right-hand side of formula (22) is then a periodic curve-factor, and serves to define the field of flow or induction outside the boundary curve. Attention being directed to such a smooth curved boundary, the angle (measured in the clockwise sense) which a tangent to the curve makes with a fixed direction may be called y, and dy/m takes the place of 1 - (A/r) as index to the periodic corner-factor sin {7 (w — a)/X}. Here a is a real variable which is to be regarded as varying continuously round the curve, increasing by A with each complete description of the curve in the clockwise sense. The transformation then takes the form dz = KGdw, where G = lim IT jsin 5 (w — a) ax/m d 8 T = exp | log sin ti (w - a}, (30) and, corresponding to formula (25), feos (27ra/A) dy = 0, fsin (27ra/X) dy = 0, (31) the integrals being taken over a range A of the variable a. @ is a periodic curve-factor. These formulae are indefinite until a functional relation is known or assumed between a and y, say yx =/(a). With such a relation postulated, and with the range of values of a specified as being from a =a toa=a+A, the formulae take the definite shape :— a+aA Ces = exp | log sin i (w - 0) T (a) da, (32) |, es @ma)d) daa [sin (ma f@da= 0 (3) 48 Proceedings of the Royal Irish Academy. The function f(a) must be such as to conform to certain conditions. It is clear that /’ (a) must (for real values of a) be periodic in a of period A, and it is also clear that F(a+A) —-f (a) = 27. (34) The electrostatic interpretation of the (2, w) transformation is in terms of an electrified cylindrical conductor alone in its own field; and it is known that when there is only one conductor the charge is of the same sign at all points of its surface. Hence if ds be an element of are of the boundary the sign of da/ds is everywhere the same, so that dy/da and dy/ds have every- where the same or everywhere opposite signs. Thus if the curve be every- where convex, /’(a) must be always of one sign, say positive. But if the curve may have concave parts, /’ (a) is not so restricted. Formula (34) indicates that the mean value of /’(a) must be 27/A. One way of summing up the requirements of the function /’(a) is to say that it is capable of being represented by a Fourier series corresponding to a wave-length A, that the absolute term in the series is 277/A, and that there are no terms in sin(27a/A) or cos (27a/A). Thus, for example, 2m/X + ¢ cos (47ra/A) is a possible form of f(a). A particular kind of geometrical consideration may be useful in suggesting possible forms of f(a). Consider any closed plane curve (not to be confused with the boundary curve in the z plane), whose tangent makes an angle with a fixed direction in its plane. Let ds be the element of arc; then it is known that ds is. 2 ~ cosw dw = 0, ga sinw dw = 0, (35) dw) dw for a range of w of extent 27. Let a new variable a be defined by the rela- tion w= 2a/A. The radius of curvature ds/dw is a function of w; let its form be ds/dw = \f'(Xw/27/). (36) This defines a function f(a) which is periodic and, in virtue of the relations (35), satisfies the conditions (53). In order to satisfy the condition (34) it is only necessary to choose the linear dimensions of the curve so that the peri- meter shall be 47°. For example, in the ellipse (a, 5) ds/dw x (a* cos? w + b' sin? w) > and therefore 19 A (a? cos*(27a/X) + B? sin*(27a/d)} 7? (37) is a possible form of f(a), provided the constant A be suitably adjusted. LreatHem— Periodic Conformal Curve- Factors and Corner- Factors. 49 A corresponding specification of f(a) is that if the are of any smooth closed curve of perimeter 47° be 27F(w), then F(27a/X) is a possible form of /(a) 11. While there is no general formula for expressing the integral a+r Gz | = lors sin} (= 2) | re) ae (38) which occurs under the exponential operator in formula (32), as a function of w, it will be shown that, for certain types of /’(a), the integral may be evaluated by a method of contour integration. The first step is to indicate any selected value of w, as it appears explicitly in the above formula, by ~,, and to replace the real variable a by the complex variable w, which is to be the variable of integration ; it is to be understood that when w is real it is to be the same as a. The change in the argument of f” gives a function /(w) which is identical with f(a) when w is real, but which is otherwise a function of a complex variable, possibly possessed of singularities which are quite foreign to f(a). The integral which comes up for consideration is 1 o (a ) = |— logs — (mM -— Ww “(w) dw, 39 SE = |= log sin) (om — ve) | fC) (39) and the value of this, when taken round a suitable contour in the w plane, has to be examined. The contour found to be most suitable consists, in the main, of a rectangle whose sides are in the lines p=, p=at+r, W=0, WH=e, where ¢ is positive and may be made indefinitely great. There must, however, be cuts from the boundary to infinitesimal circular cavities round all branch points and infinities of the subject of integration, and it is convenient to take for these cuts straight lines which start from the line w= 7 and run parallel to the line ¢=0. The complete contour includes each side of each cut, and the circumference of each infinitesimal circle. The point w, is taken inside the rectangle, and at this point the function m log sin {mA (w, — w)} has a branch point. For w—w, small the singularity is sufficiently repre- sented by mlog(w,-w), and it is seen that, if the point w describes an infinitesimal circle round w, in the conventionally positive sense, a constant 21 is added to the value of the function for each complete circuit. The same is therefore also true for any other circuit round w,, whatever its size or shape, provided it does not surround the other branch-points corresponding to w, + 7X, 50 Proceedings of the Royal Irish Academy. where 7 is any integer. It follows that the subject of integration in 4 has values which, at corresponding points on opposite sides of the cut to this par- ticular singularity, differ by 2if(w) ; thus the integrais #% taken along the two sides of the cut combine to give the value wo 2i| Jw) dw, or 2%} fiw.) - f(g, + tt)}. (40) ¢o+it The integral for the circumference of small radius « round w, is of the order of magnitude e log e, (it being supposed that w, is not taken coincident with any infinity of /’(w)}, and this tends to the limit zero. If f’(w) had a logarithmic singularity, it would be dealt with in a similar manner, and would yield a term proportional to the integral of log sin {7A (w - v)} along the corresponding straight cut. A simple pole of /’(w), say at W=W = gi + th, would correspond to f(w) taking the form Pw = uv)" for w- wy infinitesimally small, P being a constant. A complete circuit round 7, leaves the value of the function unaltered, so the integrals along the two sides of the cut cancel one another. This is equally true for infinities of higher integral order. But infinities of fractional order are branch-points, and for such the integrals along the two sides of the cut do not cancel one another, but give an integral which is not generally more susceptible of direct evaluation than the integral # For this reason the method of contour integration is not likely to be helpful when J’(w) has branch-points. The integral along the circumference round 1, of infinitesimal radius e, taken in the conventionally negative sense, is readily seen to have the limit Ficurr 3. - 2iP log sin {7X1 (uv — v)}. The treatment of an infinity of higher integral order is sufficiently illustrated by considering the case in which f’(w), for w — v, small, tends to the form QO(w—-wy?+P(w-u)". Near v,, approximately, 1 _ 7 1 . “ww 7 a log sin = (wv — wv) =— log sin x (w,- UV) - X ~ cot x (wv -— “%); A 7 Learunm— Periodic Conformal Curve- Factors and Corner-Factors. 51 and therefore the subject of integration in ¥% is, to the same degree of approximation, “ew — w,)* log sin = (wy- Ur) +(w-w,) 4 E log sin 5 (0% =W,) - " eot 5 (w, - w) For the small circular contour, described in the negative sense, the first term integrates to zero, and the second gives ag X cot Cr = w) |, (41) — 20 [? log sin = (w, =) - which is therefore the limit for a circle of vanishing radius. In considering the contribution made to the contour integral by the sides of the rectangle which are parallel to @ = 0, it is to be noticed that f’(w), being periodic, has the same value at points on the two sides corresponding to the same value of y, no ambiguity of value being possible in the region defined by the complete contour. It is also to be noticed that at such corresponding points sinwA7(w,-w) has values which differ only as to the factor — 1, so that the logarithms differ only by the constant dm ; in fact, for wy < y,, the imaginary part of ma !log sin {rA7!(w,-w)} is im 1 tan! [tanh {aA7* (Wo — )} cot {7A-1(, — )} |, which increases by 7 as # increases by A, being kept constant. From this it appears that the algebraic sum of the integrals along the sides p=a, p=a+A, is a+ it i| f/(w)dw, or i{f(a+it)-f(a)}. (42) In considering, lastly, the integral along the line w =¢, it is to be remembered that the subject of integration has a discontinuity owing to the fact that, at the beginning of the cut to w,, m ‘log sin {7A (w, — w)} is less by 2¢ for ¢ just less than ¢, than for just greater than ¢,. If con- tinuity be restored to the logarithm, for the purpose of this particular integ- ration, by continuing its analytical form from that for ¢>9,, the integral along this side of the rectangle is at+it at+it | m log sin {7A} (ww - w)} f'(w) dw - 20} f’(w) dw. (43) ataA+it go + it All singularities having been kept outside the contour, the complete con- tour integral must be zero. On combination of formulae (38), (40), (41), (42), 8 K.T.A, PROC., VOL. KXXIIL, SECT. A. 52 Proceedings of the Royal Irish Academy. and (43), the vanishing of 7 is expressed by the equality J =2Z[Plogsin {zA+ (wv, - w)} - wA* Q cot [7A™ (wy — te)} + ete. ] +ttf(a)+f(@+ tt) — 2f(w,)} a+ir+it +| wlog sin {7A7 (a, — w)} f’(w) di, (44) ~ atit which, when vv is substituted for w, after integration, constitutes an expres- sion for fas a function of w. The limit of the right-hand side for ¢ > + @ may be a comparatively simple form. With a view to investigating this limit, it can be verified that, for J great and positive, max * log sin = (wv, —w) == log (- 3i) + ae Ww) -= SS * exp| m=1 me (w-w,) ¢- (45) Let it be assumed that 7/’(w) is such a function that it can be represented, for great positive values of y, by the series io F'(w) = 3 ysexp [- 271A" (n - 5) v}, (46) where z is an integer: and let it be assumed that this series is integrable, so that, for ~ great and positive, . 2="-1 s=20 J (w) = yn + = E saps | Ie = exp [= 22th '(n — 8) w}. (47) sonti) 1 — As {7'(w)dw, taken round the rectangular contour, equals £27iP, to which integral the contributions of the sides ¢ constant, and y zero, are respectively zero and 27, it is necessary that 27 —y,A = 22iP, so that yn = 2x7 (1 — SP). (48) With the above assumptions, for y great, a log sin {7A (wy—w)} f’(w) = B+ C+ F(w,) + iA7(wy - wv) fw), where £ cousists of exponential terms, ( is a constant, and =n-1 F(w,)=-7 > i {— 2riA-!(n —s) ,}. (49) N—S LearHem— Periodic Conformal Curve- Factors and Corner-Factors. 58 Consequently at+A+it at+aAt+itt | 7 * log sin {wA-* (w, — w)} f’(w) dw = CA + AF(w,)=-107 | (w-w,) f’'(w) dw at+it a+it at+Atit atAtit = E iA1(w = wf) | +1A1| f(w)dw+Crh+rAF(w,). (50) a+it a+it In the last integral f(w) may be replaced by ynw without altering the result of the integration, and in the immediately preceding expression it is to be noted that f(a@+A+ zt) = yArA+f(a+ i). Thus the whole expression reduces to CK = Uf (a + rt) + VynWy — S4yndrA + AF (w,). (51) The substitution of this in formula (44) leads to J = 213(P log sin{wA(w - w;)} - rA1Q cot{rA1(w — w,)} + ete. ] — 2if(w) + 2m1(L -— WP) w+ AF (w) + C, (52) where w has been substituted for w,, and C” represents a constant. In this expression it is interesting to note how the infinities under the first = sign cancel the infinities of f(w) at definite points of the w plane, and. how those exponential terms of A#\w) which become infinite for ~—> +o cancel the corresponding terms in — 27f(w); thus ff has no infinities except an infinity for ~— +c corresponding to a term which is linear in w, namely — 2tmA71w. 12. Some particular cases may be considered. (i) Let /(a) be a constant, namely equal to 27A7, so that f(w) is 27A1w. Then most of the terms in formula (52) vanish, and / differs only by a constant from - 2ri\"'w; so Ge or exp J& is proportional to exp (— 2mth-'w). which is 6%. (ii) Let f(a) = 277A sinh 28/{cosh 26 — cos (47A1a)}, so that SJ (w) = - $2 log {sin (27d w — 73) /sin (27Aw + 23)}. In a strip of breadth A, f’(w) has two simple poles, namely (if a be chosen within suitable limits) at w= w, = iAB/2m and w = w, = 4A + 1AB/27, and at each of these P=-42. Thus 1 - 27P=0, and the expansion of f(w) for Ww great and positive has no term linear in w. As the expansion of /’ (w) for ~ great and positive has no term of a higher order of magnitude than [8*] 54 Proceedings of the Royal Irish Academy. exp (4ziw/X), the function F(z) is zero. Thus formula (52) gives J = C’ + log sin (w#A tw — $73) + log sin (7Atw — 478 - 37) — log{sin (27A ‘wv — 7B)/sin (27Aw + 7B)} = C” + log sin (27d + 7). The corresponding form of @, or exp / is proportional to sin (27A“*w + 73), which is @,.. (iii) Let f(a) = cos{2nzA7(a + x)}, where m is an integer and x a constant, so that f(w) =(A/2n7)sin {2nz7A "(uw +«)}. Here it is to be observed that the mean value of f(a) over a range A is not 27+, but is zero. This renders formula (48) inapplicable; in fact, y, =— 272A7=P, and the coefficient of 7 in formula (52), must be correspondingly modified. In this instance 7(w) has no infinities at definite points in the strip. But the exponential expression for # (w) contains one term which becomes infinite for ) > +o, namely dexp {— 2nizA"(w + «)j, and so F (w) = — (4/2a7) exp | — 2nizA7 (wv + «)}- Thus formula (52) assumes, in this instance, the form J = — 2 (A/2nz) sin | 2nwA*(w + K)} — (A/2nz) exp {- 2niwA7 (w+ K)} + C” = C’ — (A/2nzx) exp {2niwzA (1 + «)}, and the corresponding form of exp / is Ex = exp [— (A/2nz) exp {2nizA7(w + «)} J. (53) This is not a periodic curve-factor of the kind which has been aimed at, since it is based upon a form of f(a) whose mean value is zero, and tends to a definite limit for £2-—>+c instead of becoming infinite of exponential order 27/A. It is an inflexional periodic curve-factor whose angular period is zero. &-5 is, nevertheless, useful for the building up of periodic curve-factors of angular period 27; for, if f(a) be taken of the form 27A7 + c cos {2nwA“ (a+ «)}, the corresponding curve-factor is Gn = 6s Cr, (54) and it is clear that /.,, or any other product of ¢,, with powers of different particular cases of ¢;,, is a periodic curve-factor whose angular period is 27. _ On this may be founded a general formula. For (a) can be expressed as a Fourier series, and to each term of this series there corresponds a curve- Leatarm—Periodie Conformal Curve-Factors and Corner-Factors. 5d factor of the type of G,, except the constant term which leads to G,. The product of all these curve-factors, if it is convergent, is the curve-factor corresponding to f’(a). It may be noticed, in passing, that the form of f/’(w) suggested by formula (37), as it has branch-points, does not admit of the method of deriving the corresponding curve-factor which has just been described. 13. The specification of fields with logarithmic singularities in the region outside « cylindrical or prismatic boundary. When it is desired, particularly with a view to physical applications, to investigate fields with assigned logarithmic singularities, such as sources, vortices, electrodes, or line electric charges, in doubly connected regions of the kind under consideration, a simple formulation is available. The procedure is simply to employ periodic curve-factors and corner-factors, or any other available method, to represent the doubly connected region in the z plane conformally and repeatedly on an infinite succession of semi-infinite strips in the principal half-plane of a variable €; when this has been done, sources or vortices (in the hydrodynamical application) may be taken account of in a (w, Z) transformation, each source or vortex at a point a+78 in the € plane being accompanied by such an image at the point ¢ = a - 73 as is required to maintain the constancy of ~ along the axis of Z real. But a single source or vortex at € = Z, = a + 73 is not, in the circumstances, a representation of a physical possibility; what is wanted is an endless series of sources or vortices, all similar to one another, at the points Z = Z, + nA, where n takes all integral values, balanced by the corresponding series of images. This is the only way of ensuring that every strip in the ¢ plane which corresponds to the complete field in the z plane is equipped with a singularity representative of the single source or vortex which is present in that field. The w which, in the absence of a boundary, would correspond to such a periodic singularity, would be proportional to log [(6 — &) TE\(S - G0)? — n*A*}], or (as in article 7 above) to log sin {7A (Z — Z,)}. bs (55) When yf is to be zero for Z real, a corresponding term involving the complex conjugate to Z, must be included. Thus, for a single source at =a + 7, which produces liquid at the rate m, the form of w is w = —(m/27m) log [sin {wA+(Z — a —78)} sin {w7A71(Z- a+78)}]; (56) 56 Proceedings of the Royal Irish Academy. and, for a single vortex round which the circulation is pu, ye HE top Sa twa Ea = 28) | Qe © sin {wA7(Z — a + 73)} (57) For a number of sources or vortices « is formed by addition of terms such as (56) and (57). The elimination of @ between the (w,@) transformation formed in this manner, and that (z, 2) transformation which represents the relevant region of the z plane upon an endless series of strips in the positive half-plane of Z, gives a relation between z and w which specifies flow with the prescribed sources and vortices and the known or prescribed boundary. For example, it is known that the transformation Z = (2A/2z7) log (z/a) (58) represents the region outside the circle of radius @ in the z plane, whose centre is at the origin, upon the infinite series of semi-infinite strips of width A in the half-plane on the positive side of the real axis of Z. Hence this relation, combined with (57), defines a relation between w and 2 corre- sponding to a vortex in presence of a circular internal boundary, or to a line charge in presence of a circular conductor. The result of eliminating Z is 2 iu A sin rA™| (7A/27) log 2/a) - a — i} (59) = — og - if 27. * sin rd { (2/277) log (z/a) - a + 73} It is easy to verify that this corresponds to the familiar formula w = (ip/2r) log { (3 - 2,)/(s — &:)}, (60) where =, and 2, are image points with respect to the circle. A single vortex in presence of an elliptic boundary is represented by formula (57) in combination with s = ecosh{a — (27i/A) Z}; (61) and a single line-charge in presence of a prismatic conductor is represented by formula (57) in combination with ds/d@ = KM {sin rr(Z - y)}1~ 47, (62) the parameters being subject to the conditions explained in Article 8. These examples have not allowed for a circulation round, or total charge upon, the boundary itself, but this is easily provided for by introducing a linear term into the (rc, Z) transformation. The general form of this LeatHemM—Periodie Conformal Curve-Factors and Oorner- Factors. 57 transformation is ° Tw SO aes ee Ce sin (C —a— 73) = AMG SS log 477 sin (f- a+ 78) m ( -> aS log sin = (f - a’ - 73’) sin Fie e apy), (6 and terms representing doublets could be introduced if desirable. The formula (63), coupled with a (z, 2) transformation, whether of the type z=/(Z) or ofthe type dz=6(Z) dé, gives the specification of the field with assigned singularities in the region bounded internally in accord- ance with the latter transformation. The only further steps requisite for explicit formulation are (possibly) integration, and (certainly) elimination, and adjustment of parameters. The inner boundary may be a rectilineal polygon, in which case ((£) is a product of corner-factors in & or it may be a smooth curve, in which case G(Z) may be one of the periodic curve-factors considered above, or 7(@) may have one of the forms which have been shown above to be suitable. The limitation of the method is that one cannot prescribe the boundary arbitrarily and be sure of getting a solution ; one must be content with such boundaries as correspond to known forms of G orf, If the range of known forms of periodic curve-factors can be extended, the scope of the method will be correspondingly enlarged. Meanwhile it is possible that a rough approxi- mation to any particular assigned boundary might be got by a suitable choice from among the focal eurve-factors. R.I.A. PROC., VOL. XXXIII., SECT, A. [9] V. THE LARGE IONS AND CONDENSATION-NUCLEI FROM FLAMES. By H. KENNEDY, MA., MSc. University College, Dublin. Read Jusz 26. Published Ocronxx 3, 1916. Iv previous papers.’ Professor M-Clelland and the author have given the results of observations on the large ions occurring in the atmosphere. These investigations included a prolonged study under various conditions in the city, and an examination of the air at a distance from the city, and free from the artificial sources of large ions, such as flames, &c. The results were such as to lead to the conclusion that the large ions occurring in the atmosphere of a city are due for the most part, if not entirely, to the great number of sources of combustion, which, as laboratory experiments prove, produce large ions of exactly the same mobility as those occurring in the atmosphere. Attention was also directed to the analogy existing between the results of the study of atmospheric large ions and the work of Aitken on condensation-nuclei in the atmosphere under various conditions, and it was suggested that the nuclei measured by Aitken were not dust particles in the form of solid matter ina very fine state of division, but were identical with the large ions and the uncharged nuclei, from which large ions may be formed by ionizing the air in which these nuclei occur. The atmospheric large ions and condensation- nuclei, then, seem to be the same as those produced by fames. The mobility of the large ions from flames has been already investigated. Its value is found to be about 0003 cms. per second for an electric field of a volt per cm., and the mobilities of all the ions are the same. If the flame-gas be deionized by an electric field when the gas has just left the flame, nuclei are still formed, and they may be changed into large ions of the same mobility by ionizing the air in which they are contained. The relations existing between the number of charged and uncharged nuclei at any time, and the nature and cause of the disappearance of ions and nuclei, seemed to be a subject worthy of investigation, and the present paper is an account ? M‘Clelland and Kennedy: Proc. Royal Irish Academy, vol. xxx, Sect. A, No. 5. Kennedy: Proc. Royal Irish Academy, vol. xxxii, Sect. A, No. 1. Kennepy—Large Tons and Condensation- Nuclei from Flames. 59 of experiments carried out with the object of obtaining more information on the nature of the large ion and the nucleus from which it is formed. The general method of working was to determine the number of condensation- nuclei by Aitken’s apparatus, and to measure the quantity of ionization by an electrometer. One of the results obtained is that the charge on the large ion is not the simple electronic charge, but some multiple of it. It will be con- venient, however, to consider the quantity of electricity per c.c. of either sign as being equal to We, where « is the electronic charge. JV will not, therefore, as assumed in the previous papers, be the number of large ions of one sign per c.c. of the gas. Disappearance of the Large Ions. In experiments on the variation of the ionization with time, a gasometer of about 450,000 cc. capacity was rapidly filled with the gas from a Bunsen flame burning under the funnel-shaped vessel, illustrated in fig. 1. When 7 Z =» To Gasorneter Fig, 1. filled, the gasometer is connected to the measuring apparatus, which consists of a brass tube A (fig. 2), and an insulated concentric rod &, connected with Se ieee Fig. 2. the electrometer. An air-current of about 60 e.c. per second is sent from the gasometer through A, which is charged to about 600 volts, sufficient to remove all the ions from the air-stream. Measurements were taken at intervals for about three hours after filling the gasometer, and the values of WV deduced, assuming for « the value 4:77 x 10° E.S. units, The result of such an experiment is given in the following table :— (9*] 60 Proceedings of the Royal Irish Academy. TABLE I. Time (minutes) N x 10% Vilas! OS 0 1045 96 7 767 130 17 585 171 27 469 213 38 411 243 50 380 263 63 271 369 80 240 416 83 240 416 105 199 502 107 200 500 127 169 592 130 169 592 154 138 725 157 130 769 169 122 820 172 118 847 and is represented graphically in fig. 3. 1100 1000 909 800 700 10m x 600 N 500 206 100 It will be seen that the rate of — 60 80 100 —— 180 120 140 =160 Minutes Fie. 3. Kunnupy—Large lons and Condensation- Nuclei from Flames. 61 disappearance is very much slower than in the ease of small ions; at the end of three hours, there still remains a considerable quantity of ionization. Fig. 4 shows the graph of V' and the time. The graph is a straight line, showing that ike cal Udder ptanreny/Y IM IK Bt, IV, being the value of WV at the beginning, and (3 a constant. The large ions, therefore, recombine according to the law dN 5 vi =S—> BN . Aa 190a a | ° 10) 300 200 (0) 20 40 60 80 loo «36 120—Ss—s«st'40.—Cté=«~SSC—S—SC=idB| Minutes Fic. 4. It will be seen, however, that towards the end the rate of disappearance is more rapid than the law indicates. The value of (3, the constant of recombination, is 6°3 x 10°, taking the second as the unit of time. 62 Proceedings of the Royal Irish Academy. Measurements of the Condensation-Nuclet. The number of nuclei per c.c. of the gas at any time was obtained by means of Aitken’s counting apparatus described in detail in the “Proceedings of the Royal Society of Edinburgh,” vol. 18, z F pp. 39-52. As this form of the apparatus, however, is suitable only for measurements in air in which there are comparatively few nuclei —less than 10,000 per e.c.—and as the num- bers per c.c. in the flame-gas were as high as 1,000,000, it was necessary to dilute the gas containing the nuclei with a known volume of filtered air so as to get a concentration suitable for measurement with the counter. The dilu- tion was made in the apparatus shown in fig. 5. A is a glass vessel of about 250 c.c. capacity fitted with taps at Band C. JD is a three-way tap by means of which the vessel A may be put in communication with the 7-piece F or the tube Z, which itself is fitted with a three- way tap G, leading either to the Aitken appa- ratus or to a gasometer filled with thoroughly filtered air. Below the tap B there is a flexible tube attached, by means of which the vessel 4 is connected to a cistern containing mercury. By raising this cistern the whole vessel may be filled with mercury up to F. At K there is a short length of rubber tubing. The gas to be tested flows from the gasometer through the tube HF. In making an experiment the tap D is turned to give communication be- tween A and F, and the whole vessel is filled with mercury up to F. The merenry cistern is then lowered until there is a known volume, say 10 c.c., of the gas below D in the tube DC, which is graduated. The tap D is then turned round to connect A and Z, and the tap @ is turned so as to connect with the gasometer containing the filtered air. The filtered air is allowed into A until a certain volume, say 200 cc., occupies the vessel below C. The taps B and @ are then closed, and, by making the remaining Kennepy— Large Ions und Condensation-Nuclei from Flames. 63 mercury move to and froin A without splashing, the filtered air is thoroughly mixed with that containing the nuclei, and the dilution is such that the nuclei have only ;},th of their original concentration. By repeating the process a dilution of +}, is obtained, and, similarly, any intermediate amount of dilution may be produced. When the suitable concentration has been obtained, the vessel A is put in communication with the Aitken apparatus, and a portion of the air in it is passed through the apparatus so as to drive out the air in the intervening tubes. In this way all the space joining the counting apparatus to the vessel A is filled with air at the required concentration, and the test is made as described in Aitken’s paper. About ten tests were made of each sample of air, and the mean of these results taken to calculate the number of nuclei per ¢.c., correction being made for the variation in the pressure in A by the introduction into the counter of the amounts of air necessary for the tests. Disappearance of the Nuclet. Investigation was first made on the nature of the disappearance of uncharged nuclei, and for this purpose the gasometer was filled rapidly with flame-gas as indicated in fig. 1, but in this case the flame-gas immediately after leaving the flame was subjected to a very strong electric field to remove all the ions. The gasometer then contained great numbers of uncharged nuclei, but tests made with the electrometer showed that the number of ions present was too small to be measured. Measurements of the number of nuclei per e.c. were made at intervals, and Table II contains such a series of measurements, being the number of nuclei per c.c. TaBLe II. Time (minutes) eX alOss mt x 107 0 1125 8-9 25 306 32:7 55 164 61 965 108 92°5 64 Proceedings of the Royal Irish Acadeny. Fig. 6 shows the graph of the variation of the nuclei with time. The number per e.c. falls off very rapidly at first, but the rate of diminution diminishes with the concentration. The rate of disappearance and its nature preclude any possibility of explaining it by a simple fall under gravity. The nuclei are so small that de Broglie’ was unable to distinguish them with 100 1030 300 Minutes Fic. 6. the ultra-microscope, and consequently the amount.of their fall would be quite negligible in the time considered in the experiment. A feasible explanation, however, may be based on the coalescence of two nuclei colliding with sufficient energy, such a collision accounting for the disappearance of one nucleus. ‘De Broglie: ‘‘ Comptes Rendus,”’ vol. 148, 1909, p. 1317. Kennepy—Large Ions and Condensation- Nuclei from Flames. 65 There are many difficulties in understanding the origin of the large ion or nucleus, but Langevin’ and Sir J. J. Thomson’ have given a satisfactory theory of the nucleus in its final state as consisting of a minute water-drop retaining a stable size in unsaturated air, the size being almost independent of the charge carried. When two such nuclei come into collision and coalesce, the two will form one tiny drop with a radius greater than the equilibrium value, and evaporation will take place until equilibrium is restored, so that the new nucleus will have the same size and properties as either of the original ones. On such a theory it is to be expected that the frequency of the collisions, and consequently the rate of disappearance of the nuclei, should be proportional at any time to the square of the number present. ile Base : Fig. 7 is the graph of 7 with the time, and shows that and = == yn, n, being the initial value, and y a constant. Experiment therefore supports the idea of the disappearance of the nuclei being due to collision and°consequent coalescence. a fo It next seemed of interest to ascertain A what effect the electric charge has on the col- S lision frequency and rate of disappearance of o Be the nuclei, and for this purpose similar ex- periments were made when the gasometer = was filled with flame-gas without having re- moved the ions. In this case some of the nuclei are charged positively, an equal number 20 negatively, and the remainder uncharged. The uncharged nuclei were in such experiments 0 20. 20 2c bout half or one-third of the total number. Minutes Table III shows the result of such an Fic. 7. experiment. 1B. Bloch: ‘*Ann. de Physique et de Chimie,” 1905; Chauveau: ‘‘ Le Radium,” Avril, 1912. 2 Conduction of Electricity through Gases. EB R.1,A. PROC., VOL, XXXIII., SECT. A. [10] 66 Proceedings of the Royal Irish Academy. TABLE IIT. Time (minutes) i se Wr > 107 I | 0 829 12 20 345 29 43 219 46 74 145 69 115 103 99-5 154 76 132 223 50-4 199 Fig. 8 is the graph of 7 and the time, and fig. 9 that of : and the time, showing that in this case as well the rate of disappearance of the nuclei is 900 500 400; n x 107 300 QO 40 80 120 60 200 249 280 Minutes Fie. 8. proportional to the square of the number present. Further, as far as the experiments have gone, there seems to be no appreciable difference in the Kunnepy—Large Ions and Condensation-Nuclei from Flames: 67 rate of disappearance of the nuclei, whether they are uncharged, as in the first series of experiments, or whether, as in the second series, a great proportion of the nuclei are large ions. The rate of disappearance seems to be the same, within the range of experimental error, though of course further work is desirable in order to make this point certain. 226 200 ° (80 160 140 {20 Tis entoe {00 80 60 40 0 40 80 120 (60 200 240 280 Minutes Fie. 9. The mean of a number of determinations of y is 14 x 107°. In all cases of the disappearance of the nuclei when very long intervals are considered, the nuclei disappear more rapidly than according to the formula dn dt In one experiment the number of nuclei per c.c. at the end of twenty-one hours [10*] == yn. 68 Proceedings of the Royal Irish Academy. was 1800, having diminished from 1,000,000 in that time, whereas the value should be about 10,000 if the formula held true during that period. The Charge on the Large Ion. In some of the earlier experiments on the decay of the ions and nuclei from the same mass of flame-gas in the gasometer, it was found that in certain cases the total number of nuclei per c.c. was less than the number of large ions per c.c. as calculated on the assumption that each ion carries the electronic charge. This leads to the conclusion that the ions carry multiple charges, and an attempt was made as follows to determine the charge. The flame-gas was drawn slowly—at about 60 c.c. per second—through the funnel A and the large vessels B and C—each of about 70,000 cc. capacity—as illustrated in fig. 10, so that on leaving C’ the ions and nuclei had reached their stable size. The method adopted was to determine the total charge fo ee ) | | } || i i a ie bay bel | \ | | | A B caee| 4 | II | | Fie. 10. per c.c. by the electrometer, and to measure the total number (z) of nuclei by means of the counting-apparatus, and also the number (7) of uncharged nuclei left in the air-stream, when all the charged nuclei had been removed by an electric field. — is the total number of charged nuclei, and, as in the gas from fiames, the positive and negative charges per c.c. are equal, half n= 2 charge per c.c. measured by the electrometer, and from this the average charge may be calculated. The results of a number of such experiments, made on different occasions, are given in the following table, the positive charge per c.c. being given in the fourth column as the multiple WV of the electronic charge «. The average charge Z on the ion is given in the fifth column as a multiple of «. this number, aad may be taken as the number of nuclei carrying the a eee me | Kunnepy—Large Ions and Condensation-Nuclei from Flames. 69 TABLE IV. n x 1075 no x 10-* (n — mo) x 10-5 N x 10-5 | = 673 | 377 296 260 | 1:8 828 480 348 348 | 2-0 870 | 539 331 323 | 1-9 781 | 373 408 363 1-8 | 862 521 341 322 1:9 1006 613 393 388 2:0 836 399 437 593 | 2-7 715 339 376 499 2-7 644 358 286 450 | 3-1 768 391 Og 493 2-6 740 355 385 375 | 2°0 873 511 362 395 2-2 A series of measurements were next made in which the flame-gas was drawn into the gasometer at the same slow rate, and was stored for various time-intervals ranging from one and a half to two and a half hours before the tests were made. Table V gives the results of a series of such experiments. TABLE V. | E nm x 103 no x 10-3 (m=) x 10-8 Nx 10-3 a 202 108 94 80 1-7 82 29 53 72 2°7 75 30 pet 71 371 130 64 66 123 3:7 136 62 74 104 2°8 65 37 28 57 4 82 29 53 69 2-6 140 | 67 73 103 2:8 76 24 | 52 61 2°3 95 | 24 71 64 1-8 141 51 | 90 73 1°6 130 | 78 52 92 3°5 70 Proceedings of the Royal Irish Academy. A number of determinations of the charge was made also from the decay curves of V,n, and m, the gas in these experiments being drawn very rapidly, as in fig 1, into the gasometer, and tested at intervals for two or three hours afterwards. From the decay curves, the values of NV, n, and m are obtained at any time, and the charge calculated. In all such observations the value obtained for # was very much greater than in the experiments when the gas was drawn slowly from the flame, its magnitude varying from about 4:5 to 6e. There is no doubt, therefore, that the large ion bears a charge which is a multiple of the electronic charge. The magnitude of the charge seems to depend on the circumstances under which the ion is formed. There are several reasons for believing that the formation of the nucleus does not depend on the presence of the charge. Experiments made on the number of nuclei in the flame-gas show that the number is not diminished appreci- ably by the removal of the ions just as the gas leaves the flame. The gas is then at a very high temperature, and the molecular groupings forming the nuclei have not been formed before the ions are removed; and yet their growth is not apparently affected thereby. The charging of the nucleus seems to be an accident, not essential to its formation, but occurring owing to the enormous number of free ions of very high mobility in the flame and the gas coming from it. Considering the charging to take place in this way, it is very easy to understand the formation of multiply charged ions, and it is to be expected that the average charge on the nucleus should depend on the number of free ions of high mobility present when the nuclei are in process of formation. The much higher value of the charge ZH, obtained above from the decay curves of V,n, and 7%, might be explained by the consideration that in all these experiments the gas was drawn from the flame with great rapidity into the gasometer. The gas therefore cooled rapidly, and the nuclei were formed when there was a much greater number of free ions present than in the first series of experiments, where the cooling of the gas was slower and the nuclei had not been formed before a comparatively great number of the free ions had disappeared by recombination. An estimate of the charge on the large ion may be deduced directly from the constants of decay 3 and y of the charge per c.c. and the nuclei, assuming that the rate of disappearance of the nuclei is not affected by the charge. For, suppose the charge per c.c. is Ve, then If there are , charged nuclei of either sign, each carrying a charge equal to Kennepy—Large Ions and Condensation- Nuclei from Flames. 71 ze, then N-=xn,, and consequently dn, =-— Brn? dt Band, die ae ? F : where Gp Means the collision frequency between ions of different sign, every such collision involving the loss of charge of the two colliding nuclei, and resulting in the formation of one uncharged nucleus, with the subsequent history of which we are not concerned, since it is only the charged nuclei that are under consideration at present. If there are m nuclei per c.c., it has been found that that is, the collision frequency is yn’, each collision involving the loss of one nucleus. With equal numbers of positively and negatively charged nuclei, or, 2n,- altogether, the collision frequency will be 4yn,.?. If the collision frequency be independent of the charge, as seems true from the experiments, the number of collisions per second taking place between the positively and negatively charged nuclei will be half the entire number, or 2yn,’. Fherefore a =—2yNe’, and consequently : Bx = 2y Now GSG3 x 10, eam@l 57 = idl s< 10=8, therefore AB = flo) and this value agrees well with the values obtained under similar circum- stances from direct measurements of JV, n, and m, the calculations being made from the decay curves of these quantities. The Mobility of the Large Ion. The results given in this paper show that the charge on the large ion may have widely different values under different circumstances of production. Even in any one mass of flame-gas it is not likely that the values of the charges borne by the different ions will be the same, though they will probably vary but little round a mean value. When the gas has been deionized by an electric field, and the nuclei charged again by ionizing the 12 Proceedings of the Royal Trish Academy. gas with X-rays, the charge on the nucleus will depend on the imiensity of the ionization. Im all cases, however, as far as experiment has gone, the mobility has the same value (005. The mobility must, therefore, be inde- pendent of the charge, a result which is m agreement with Sir J. J. Thomson's! theory of the mobility of ions m gases. The Nature of the Large Ions and Condensation-Nuclet. Though Langevin’s theory gives a satisiactory explanation of the siable existence of a very minute water-drop. even im unsaturated air, such drops being the nuclei for condensation found in the atmosphere and produced by flames, there is yet very little to suggest an explanation of the origin and growth of the nucleus itsel—i It has been already stated that the charge can play no important part im its formation. since the number produced by the flame is not affected by the removal of all the free ions just as the gas leaves the flame. On the other hand, the production of nuclei in flames is associated with the production of water-vapour. Ions of the same mobility 0003, and neutral nuclei, are also produced by the burning of hydrogen: but flames such as CO, sulphur, or arsenic, in which water is not produced, give no large ions. That the presence of water-vapour, however, is not sufficient for the production of large ions and condensation-nuclei has been. demonstrated by the experiments of de Broglie* and Aitken on the hydrogen fame De Broglie, using a flame of hydrogen thoroughly dried and purified. and burning at the end of a lead tube in a chamber Kept cool by a water-bath, found that no large ions are produced. Aitken with similar precautions as to purity found that hydrogen burning in thoroughly filtered air produced no condensation nuclei, but if any solid impurities, such as particles of dust or rubber from the connecting tubes, found their way into the flame, great numbers of nuclei were immediately formed. Further experiment on the early history of the large ion, such as, for example, the variation of the mobility as the gas from the flame cools, may give more information on the nature and origin of the nucleus formmg the large ion. There are two other allied phenomena, which in this respect seem worthy of further study. It is well known that when all condensation nuclei have been removed from a mass of air, saturated with water-vapour, either by repeated expansion or by filtering, an exceedingly dense cloud is formed ' Sir J. J. Thomson: Proc. of the Physical Society, vol. 27. Dee., 1914 *de Broglie: ““Comptes Rendus,” vol. 151, 1910, p. G7- Kennepy—Large Ions and Condensation-Nuclet from Flames. 73 by producing a sufficiently great expansion and consequent cooling and supersaturation. If this cloud be evaporated, by the heating due to sudden compression, it disappears very rapidly, but a very small expansion again produces a dense cloud. It is evident, therefore, that the drops of the first cloud did not entirely evaporate, but remained in a stable state in the form of very minute invisible drops, as suggested by Langevin’s theory. These invisible drops serve as nuclei of condensation for a very small expansion, apparently similar to the nuclei produced by flames and occurring in the atmosphere. Again, de Broglie! has shown that when pieces of moist pumice are heated, nuclei are driven off from the surface, and when these nuclei are charged by ionizing the air containing them, the mobility of the ions so formed is the same as that of the ions from flames. He points out, too, that the formation of the nuclei is due, not to the production of large quantities of water-vapour, but to the driving off of the last surface layer of moisture. Before, however, being able to identify the nuclei obtained in these two ways with those produced by flames, it would be necessary to see if the manner of disappearance of the nuclei is the same as for those from flames. Summary. 1. The rate of decay of ionization in the case of large ions from flames is according to the law dq F at == Bq A where g is the charge of one sign per c.c. 2. The rate of decay of nuclei, measured by Aitken’s apparatus, is accord- ing to the law and is the same whether the nuclei are charged or uncharged. 3. The large ions de not carry the simple electronic charge but some multiple of it. The charge varies under different conditions. 4. The equations in (1) and (2) do not hold when the time gets very great, ‘de Broglie: ‘* Annf de Chimie et de Physique,” vol. 16, 1909. R.I.A. PROG., VOL. XXXII, SECT. A. (11) 74 Proceedings of the Royal Irish Academy. the rate of falling off being then in both cases greater than would be given by these equations. 5. The mobility of the large ions is the same under all circumstances, and the same in any mass of flame-gas. The mobility must, therefore, be in- dependent of the charge. 6. The formation of the nucleus seems not to be affected by the charge. In conclusion I wish to express my indebtedness to Professor McClelland for his advice and encouragement during the work. VEE IMPACT IN THREE DIMENSIONS. By PROF. M. W. J. FRY, M.A., F.T.C.D. Read June 26, 1916. Published Fepruary 9, 1917. 1. The problem of impact in three dimensions is incompletely discussed by Routh in his excellent treatise on Elementary Rigid Dynamics, of which a seventh edition appeared in 1905, and which may be regarded as the standard work on the subject. He does not show how the initial stage of the impact depends on the roots of a certain equation #'(@) = 0; nor how generally the representative point gets on to the line of no sliding; nor that when sliding ceases and rolling is impossible the motion of the representative point is along a line determined by one particular root of (0) = 0. Also, the solution he gives of impact between perfectly rough bodies is not correct, as it may involve the physical absurdity of supposing the impulsive normal reaction to be negative. These points and others are discussed in this paper, and it is shown that the course of the impact under the most general conditions in three dimensions can be minutely traced, and lastly the correct solution of the problem is given when the coefficient of friction is supposed to be very great. 2. The General Equations.—In order to deal with the problem of getting the motion after impact, when a rigid body A strikes against a body J’, we trace the variation of the resultant blow delivered by A at the point of contact O, by resolving it into three components :—P and @Q the components along any two perpendicular axes drawn through 0 in the common tangent plane, and & the component along an axis drawn in the direction of the common normal, so that it is initially and always positive. We follow the movement of the extremity of the resultant blow, whose co-ordinates are P,Q, R, and call it the representative point. For A, let WM be its mass; u,v,w the components of the velocity of its centre of gravity; @,, wy,w, the components of its angular velocity ; U,V, W, Q:, Qy, Q: the initial values of the same quantities; A, B, C its R.1,A. PROC., VOL. XXXIII., SECT, A, [12] 76 Proceedings of the Royal Irish Academy. moments of inertia; x,y,z the co-ordinates of O; Lm, Lm n., lymsns the direction cosines of the axes of P, Q, R; all referred to the principal axes through the centre of gravity of A. Let the corresponding quantities for A’ be denoted by the same letters with dots. Then as A has been acted on by a blow - P, -Q, - fF at a, y, 2, we have:— M(u — U) oe (AP oP LQ at 1,R). M(v-V) = — (mP + m,Q + mf). M(w - W) = - (mP + 2.Q + 23h). ie Y, MP + mY + mh | _ zs io 4i(Ge= Oe) = = z, mP + mQ + mR| Re = OD Rell 8, MP + mY + nsf es ee aS neal TP B(wy — Qy) Ee” LOU ER faP.— p2Q — pslt. (1) r, UP + LQ + 1k : = ="— == MoS = pli, es Cay Y, MP + mQ + mR re ees The corresponding equations for A’ are obtained by pee - P, -Q, - & for P, Q, R, and dotting all the other letters. These equations connect the velocities of translation and rotation of A and A’ with the total impulsive components P,Q, & at any stage of the impact. At any such instant, let S inclined at 6 to the axis of P be the velocity with which the point O of A slides on 4’; let K be the velocity of com- pression ; and let the initial values of these quantities be denoted by S,, 9,, K,; then we have Scos@ = 1,(w + zwy — yw) + M (v + wz — Zwz) + 1, (w+ Ywr — Lwy)—(similar quantity with dots) = hw + mv + Nw, + Wed, + Wyly + w:2,—(Similar quantity with dots) = S,cos 0, — == SOP + 2.0 +R) - (uk + + wo) 2 gue + 12@ + vR) iP Xr 4 ; / / = W = 7 (AP a5 A/Q + Aj) = = (ui P a pelQ + ps 2) = = (w'P + w'Q + v/R) S, cos 6, - aP — hQ - gh. Fry—Impact in Three Dimensions. Ue Similarly we get Ssin 6 = S, sin 0 - AP - 6Q - fR, (2) ie = UG = GIP = 9 =o ee naa? Bites in NO ae a pal? + 2 MM” Ww A B Cord Bi (OF poe igre ge Ee eee with corresponding values of 6,¢,g,h, so that it is easily seen that a, b,<¢, be-f?, ca-g, ab-h?, and A = abe + 2fgh - af’ — bg’ —-ch* are all positive. 3. During the course of the impact an instant ought to arrive when XK, the velocity of compression, vanishes. If &; be the value of / at that instant, it is an experimental law that the impact is over when & becomes equal to (1 + e)F;, where e is the coefficient of restitution. In two dimensions & vanishes only once during the impact, but in three dimensions I find that A may vanish once or thrice. This result no doubt conflicts with our preconceived ideas about impact, but so do other results which undoubtedly hold in two dimensions also, for instance: —S may begin by increasing, and K also may begin by increasing. When X vanishes three times, and /¢ does not attain the value (1 + e) Ri while K is negative, that is between the first and second vanishing of Jf, then we must take R; to be the value of & when /f vanishes for the third time, so that /f shall be negative when £ attains the value (1+e) A; Between the second and third vanishing of &, is positive, and the impact could not be over when the bodies are still compressing each other. 4. When the bodies are perfectly smooth, P= 0, @=0 during the impact, so that &; = I,/c, which is positive, as Kj and ¢ are both positive. The final values of the velocities of rotation and translation are then obtained by puttmg P=0, Q=0, R=(1+e)A; in the equations (1). 5. If the coefficient of friction mu is very large, it is commonly supposed that when K = 0, S=0 also, so that AR; = (ab — h*) K, + (hf — bg) 8, cos 6, + (gh - af) S, sin 0. If, however, we treat the problem in this way, as K,, S,, 0, may have any values provided &, and S, are positive, we can arrange an impact such that &; is negative, which is absolutely impossible. For instance, taking 0, and 6) +7, the factor multiplying S, changes sign, so that if (ab-h*) K, is taken less than S, multiplied by the absolute value of that factor, Ko, S,, and either @ or 0,+7 give an impact for which /; is negative. This proves [12°] 78 Proceedings of the Royal Irish Academy. that the commonly accepted solution of the problem of collision between very rough bodies is unsound. I give the correct solution atthe end of this paper. It is not very much less complicated than the general solution. 6. Continuation of the general theory. In the equations (2) we take S to be positive, its direction being given by 9. It is most important also to keep in mind that & increases continually during the impact, so that it forms a useful independent variable in terms of which to express S and 0. Differentiating equations (2) d (Scos 0) = — adP —hdQ - gdh d (S sin 8) = — hdP — bdQ - cdR. Now, while sliding is taking place, the increment of the impulsive force of friction on A’ is in the direction of sliding and equal to pdR: dP =n cos @dk, dQ = sin 04k; 1S ; dé : cos I ~ Ssin 6 = =-(ancos0+husin§+g)=- VU, : aS 0 . : sin 6 5 + S cos 6 — = — (hu cos 6 + busin @ + f)=— U2; -, a0 Sie St U,sin@-U,cos0= F(@) as L ap 77 Uic08 8 - Uz sin 6 = — $(9). 7. If initially #(@,) = 0, and S is not zero, @ will remain constant, and therefore the representative point will move along a straight line inclined at tan‘ to the axis of &. For, as £'(8,) = 0, and S is not zero, = = 0 initially. Differentiating _ a6 x ds dé dF (8) dO CP GR ane Bakar: ha) i apa 0 initially. Similarly all the derived functions of 6 with respect to & vanish initially, so that 6 remains constant. There are two or four values of 6 which make #'(6) = 0; for by putting z=peos$, y=psin8 in F(@), we see that such values of @ are given by the intersection of the rectangular hyperbola (a —b) ay+h(w-2#) + gy -fe = 0, which we shall call H = 0 with the circle z* + 7* = y, whose centre is on the Fry—IJmpact in Three Dimensions. 79 curve ; so that there are always two real points of intersection, and four if u is large enough to make the circle cut the other branch of the hyperbola ; this value of « we shall call jp. If at any instant during the impact @ should become equal to one of the roots of #'(@) = 0, we should expect the representative point to move as described in this section; but as a matter of fact the conditions set down here never occur unless they do so initially, for we shall see that whenever 6 becomes equal to a root of (0) = 0, at the same time S = 0, so that we are brought to the consideration of what will happen when S = 0 initially or at any moment during the impact. This discussion of the roots will be further developed and also another discussion given in section 9; both of which will show that when p is less than the value jy, necessary to make rolling possible, one root exists for R as is positive, and one or three other roots exist for which ¢ (6) is positive ; and that when p is greater than ju, ¢ (0) is positive for the two or four roots which then. exist, so that when sliding takes places initially along any of these other directions, S decreases continually. which # (6) is negative, so that when sliding is along it S increases as 8. To simplify further discussion we shall show (see Routh’s Rigid Dynamics) that by turning the axes of P and Y about the axis of #& through an angle y, / may be made zero. For after the axes are turned through y, let the values of P, @, S, 0 for the new axes be denoted by the same letters with dots, and we get S’cos 0 = Scos @ cosy + Ssin @sin y = (S,cos 4, — aP - hQ — gR) cosy + (Sp sin @ - hP - 6Q - fR) sin y S, cos 0, - (acosy + hsin y) (P’cosy — Qsin y) - (heosy + bsin y) (P’siny + Y’cos y) - (geosy + fsin y) & = Socos 8,’ — (a cos*y + 2h cosy sin y + bsin*y) P” N= \ - (= sin 2y + hcos 2y | WY -(gcosy + fsiny) R similarly ys a fis , b ee. . S’sin 0’ = S,sin 0,’ — (+ sin 2y + 4 cos 2y) P — (asin*y — 2h sin y cos y + 6 cos*y) YY — (feos y — gsin y) R. Thus h = 0 if y has any of the four perpendicular directions given by 2h a-—b tan 2y = 80 Proceedings of the Royal Irish Academy. I further note that we may take @ greater than }, for if a is less than 8, interchange the axes of P and Q, also that we may take f and g to be each negative, for if not by reversing the direction of the axis of P we change the sign of g, and by reversing the direction of the axis of @ we change the sign of f Thus there is one unique pair of axes for which fh = 0, a is greater than b, and f and g are each negative, but the rotation from P to Q for this pair may turn out to be clockwise or counter-clockwise. 9. If S=0 initially or at any instant during the impact, as we take S to be positive, its direction being given by 6, it must either remain zero or increase. If it remains zero, the representative point moves along the line given by the intersection of the planes aP+gRk=S8,cos#, 60+ fh = S)sin &, h being taken =0. Along this line adP+gdR=0, bdQ+/fdk = 0, and as p~*(d/t)? must be greater than or equal to (dP)* + (dQ)?, motion along this line, called the line of no sliding, is possible only if p is greater than or equal to w., where ig? ae = J+. rt iE Gb If S increases from zero, as dd a Sar = F(6), =-— »(6), initially 2°(@)=0, and by ils 5 St dé dS dé dF(#) dé “aR * dR dR 40 dR- Now, i | F'(8) = z ae = U, sin - U,cos@) =U, cos@ + Uz, sin @ — aw sin’ @ — bu cos? = ¢(@) — p(@), where p(O) is positive for all values of 0, ea) (5 ee | dé dk * \" * PO) aR = ’ ReneS dS ‘ “iis LM ret as S = 0 initially, and aR and p(@) are each positive, aR = 0 initially ; differentiating again a0 (., aS (vO aos ae 2M i Ned Oar ape + pO) (55) = 0; oe = 0 initially, and similarly it may he shown that all the derived AR functions of @ with respect to & vanish initially, so that @ remains constant. Fry—Impaet in Three Dimensions. 81 Hence, when S=0, the representative point may move along a line inclined at tan? to the axis of R, whose projection on &=0 makes an angle @ with the axis of P, which satisfies /(@) = 0, provided that this value of @ makes ¢({@) negative. We shall now show that there is such a unique root of F(@)=0, if w is less than m,; so that the possibility of moving along this line and that of moving along the line of no sliding are mutually exclusive. A root @ of F(@) = 0, gives 0, U, U, cos 6 + U. sin 8 cos@ sind cos? @ + sin? 6 : (qu -)cos8+g=0, (du - o)sind+f = 0, Gf We (au — o) i (ou - ¢)? Tracing the curve ff i = : ~ + - 1, 7 (au =F * Gu = 2) a OF oF dz) G@Qi=zy (Giz)? y then increases from -1 at r=-c to +c when w=,0, assuming a greater than J; it then diminishes from + until a. —- x x—bdu i F af® + bg’ g fi fi + g ¥ at which point 9 OE EGY UR GID OGY (a-b gt (a -b) fw (a — by It then imcreases to +o for x = wa, and then diminishes to — 1 for x=+o. Thus there are always two values of # for which y = 0, one less than yd, and the other greater than ya, and there will be two other values between wd and wa if y is negative at the point where it is a minimum, or if (eee (a — by tgs This then is the value of w referred to in section 7, such that if pu is greater than p. the circle a + y* = uw? cuts the other branch of the pw’ is greater than rectangular hyperbola H = 0. Of the two or four real roots one only can be negative, and the condition that there should be one negative root is that y should be positive when z = 0, or that p2 should be less than gm Be 82 Proceedings of the Royal Irish Academy. We thus have shown that when S=0, according as pw is greater or less than s, the representative point moves along the line of no sliding, or along a line inclined at tan? to the axis of R, whose projection on & = 0 makes with the axis of P the angle @ associated with the negative value of ¢. The value of 42 can also be obtained by noting that as the circle touches the hyperbola for that value of u, we have F(@) = 0 an ao = 0, so that a-b —3 msin 20 + g sin 8 - feos@=0, (a-6) pcos 20+ 4cos8+fsin@ = Solving for f and g J = (a-b)psin? 6, g = (6-4) pcos’ 8, (a — byw? = (fF + gi) 10. aie now that S, is not zero, and that @ is not a root of F(@) = 0, sliding takes place initially, and we have dS aR = — U,cos@-U,sin@ = - o(6), 317 _ Ui, sind — U,cos6 = FO STR = U,sin@ - VU. cos (@), 1 ds 6 Ff’ ) y Sapa oa =- TR where (8) = asin’ @ + bu cos? 0. Integrating from @, to 8, S= S.H(Go) exp (- | A) a0), F(6) = \ F(0) _, oS _ SHO) (2 27a) ‘do FQ) {F@)? “PF | =| a0 dee As dit is always positive, it follows that 6 increases continually if 7 (0) is positive, and decreases continually if /'(@,) is negative, until it becomes equal to a root a of F(@) = 0. By the above equations, S and FR both remain finite until @ nearly equals a, and it is most important to find if they remain finite when 0 = a. Putting 0=a-«, 1 aoe ae FO) F@ |i fy + hye + c.f so that integrating from a small value 7 of « to «, and then making « in- definitely small, the important part of _ ff 202 i. af ee a 2S F (0) , F(a) ~F(a) near a, Fry—Impact in Three Dimensions. 83 Thus the value of S depends on In either case K will vanish : : . UP 9 : QAO once or will not vanish, according as K for > is negative or positive. If negative, the end of the impact may be in the sliding or rolling stage; if positive, it must be in the rolling stage. Thus if Kk, - 8, (4 cos, + = sin 0). which is the value of X for 0 = =. is positive, the solution is obtained by taking S=0 when #=0, except in the peculiar case occurring in (2), in which case and in all other cases the solution depends on the solution for @ of the equation =0. Also we see that, no matter how large pw may be, sliding may not cease in certain cases by the end of the impact, and in all such cases R may be taken = 0 when we proceed to find wv w wy wy wz, ete. In this discussion pu is taken to be very large but still definite, so that P and @ vanish if R vanishes. We would have to deal with a different problem, if we could assume prominences on one surface to fit into depressions on the other, so that the two bodies interlock, and P and Q can have values when &=0. In this case if A; is negative, or the point of intersection of the line of no sliding and the plane of no compression is below the plane & = 0, by taking & = 0, and finding P and @ from S = 0, we arrive at a point on the line of no sliding for which MK is negative, and conclude that the impact is then over. Kee lair Fit AP eee te ss nit hone Sear . i i pc gna a oN) AAT SAT a TT s Wane Cae AT : tire e ia lRuh ein yu ou ay ae e ; sy 7 fa EA - he 9 AD Fe Sega Said i} x Wei only ag ace Rit 5 Lisl ahh ab MN ces < My), TA UO sole : Fy VA n it ta, alae nad ne OR own weed’ eo nut ch ce ae " 1 - iA, rindi be x 1s > = fv ’ ri O16 olinlij tld Avi Weeee of Tat, 7 anion . ald pay Onl er yi APT epee ithe obhe API Fa Pee no Fee oe heli Se hae Ste ite. fetes! cue Gitielle aN: hb Real Waele Ont. teas Pal Eee Cle (Se one! ied Uses i yt urs Gy) PPA it ait emp pire hyp tie Lie Ry sere is ; i va WE wo Miler? fee el eee nilh he Sea te ’ if e obmih ie aad, <2 eee 1) eee meen yi 5 ie OPP IL zs Tliesee te lac ee uy iu 7 “pout gy elvvi Vip ult Stn ton ower CUM) Ian a HRY TIA? oN (aay Pilea Cae eee Fels et aa ae a ih tt ho eee ‘ales wa 7 ' >» i) nr? he ee ee cer : eying Gh WY AN Ee ee es i ie PROCEEDINGS OF THE ROYAL IRISH ACADEMY VOLUME XXXIII SECTION B.—BIOLOGICAL, GEOLOGICAL, AND CHEMICAL SCIENCK. DUBLIN: HODGES, FIGGIS, & CO., LTD. LONDON: WILLIAMS & NORGATE 1916-1917 Tae AcapEmy desires it to be understood that they are not answerable for any opinion, representation of facts, or train of reasoning that may appear in any of the following Papers. The Authors of the several Essays are alone responsible for their contents. Dunit: Pemren at tHE University Press ry Ponsonry AND GIpHs. CONTENTS SECTION B.—BIOLOGICAL, GEOLOGICAL, AND CHEMICAT, SCIENCE. Axear (Joseph), M.Sc. :— Diketones derived from Diacetoresorcinol-Dimethylether, Unsaturated Ketones derived from Diaceto-orcinol, Carpenter (G. H.), M.Sc. :— The Apterygota of the Seychelles. (Plates I-X VIII), Kew (H. W.) :— A Synopsis of the False-Scorpions of Britain and Ireland: Supple- ment, : - : . ‘O’Riorpan (W. M.), M.Sc. :— See under Ryan (H.). Ryan (H.), D.Sc., and W. M. O’Rrorpan, M.Sc. :— On the tinctorial constituents of some Lichens which are used as Dyes in Ireland, Ryan (H.), D.Sc., and Phyllis Ryan, B.Sc. :— On the Condensation of Aldehydes with Ketones. iii—Benzal- dehyde with Methyl-Isopropyl-Ketone, Ryan (Phyllis), B.Sc. :— See under Ryan (H.). 71 91 105 ERRATA. | e SECTION B. p- 95, 1. 9, after glacial read acetic. yy 1. 31, for 0°6580 read 0-0580_ e i + PROCEEDINGS or THE ROYAL IRISH ACADEMY PAPERS READ BEFORE THE ACADEMY Ie THE APTERYGOTA OF THE SEYCHELLES. By PROFESSOR GEORGE H. CARPENTER, M.Sc., M.R.LA., Royal College of Science, Dublin. Puates [-XVIII. Read Frrrvuary 14. Published June 5, 1916. THE collection of wingless insects described in this paper was made as part of the work of the Percy Sladen Trust Expedition to the Indian Ocean in 1905 and subsequent years under the leadership of Professor J. Stanley Gardiner, E.B.S., of Cambridge. Many of the specimens were collected by Professor Gardiner himself, others by Mr. J. C. F. Fryer; but the greater part of the collection was obtained by Mr. Hugh Scott, of the Cambridge University Museum. He spent eight months on the Seychelles during the years 1908-9; and an interesting account of his methods of work, with descriptions of the various islands visited, and the nature of the mountain-forest regions from which most of the insects come, will be found in a paper (1910) published in the Linnean Society’s ‘l'ransactions, in which have appeared most of the results of the Sladen K’xpedition hitherto issued (Gardiner and others, 07-14). For the privilege of examining this highly interesting collection I am indebted to the kindness of Professor Gardiner, to whom and to Mr. Scott my best thanks are further due for much information willingly given, and for patience under long delays due to the pressure on my time of other work. The publication of the paper by the Royal Irish Academy during war-time has been much facilitated by a grant, which is gratefully acknowledged, from the Council of the Royal Society. It is worthy of remembrance that a former Secretary of the Academy, E. Perceval Wright; made, nearly fifty years ago, a biological expedition to the Seychelles, and described some plants from the islands in our Transactions (’71). A general account of the area in which the collection was made has been given in Professor Gardiner’s paper (06) on the Indian Ocean, and in his contributions to the Reports of the Expedition (07-14), The vast majority R.I,A. PROC., VOL. XXXIII., SECT. B, [B] 2 Proceedings of the Royal Irish Academy. of the specimens come from the granite islands of the Seychelles Archipelago in the restricted sense ; most are from the forest-clad, mountainous islands of Mahé and Silhouette, 2 much smaller number from the islands of Félicité and Praslin. Collections of Apterygota were also made on Coetivy to the south and on some of the coral islands to the south-west of the true Seychelles group—the Amirante and Farquhar, and Aldabra—the latter of which he north-west of Madagascar. A discussion on the geographical bearing of the facts of distribution of the insects is given at the end of this paper _ (pp. 48-55). The Apterygota are now generally recognized as a sub-class of the Insecta, showing a number of interesting primitive characters which afford a strong presumption in favour of the view that their universally wingless condition is to be regarded as a survival inherited from the remote ancestors of insects, and not as an adaptation to some abnormal mode of life like the parasitism of such insects as lice and fleas, whose winglessness is clearly a secondary character. On account of their inability to fly, and the wide and often discontinuous range of many genera and species compared with the curiously restricted distribution of others, the Apterygota may be regarded as specially important in faunistic studies which open up problems of ancient geography. The rich collections of these insects which have been gathered in the Seychelles and neighbouring archipelagoes promise, therefore, results of some importance. The two main orders of Apterygota which were recognized by Lubbock in his classical monograph (73), the starting-point for most English-speaking students of the group, are both well represented in the collections now described. These orders can be readily distinguished by superficial characters : — A. Feelers long, multiarticulate. Ten abdominal segments. Often eight pairs of simple abdominal appendages, c . Thysanura. B. Feelers with four to six segments. Abdominal segments six only. At most three pairs of abdominal appendages, reduced or modified, . : : 3 . , : ‘ . Collembola. Order THYSANURA. The Thysanura or “ Bristle-tails ” are well represented in the fauna of the Seychelles. Hitherto only two species of the order—Acrotelsa collaris (Fab.) and Lepidospora Braueri Esch.—appear to have been recorded from the archipelago; both of these belong to the extensive family of the Lepismidae. In the collection now described, eight species (four of them new) of this family are enumerated, besides three of the Machilidae, and one each of the Campodeidae and the Iapygidae—all of these being apparently new. ‘Thus the four principal families of the Thysanura have members among the insects Carpenrer— The Apterygota of the Seychelles. 3 of the Seychelles. These families are easily distinguished by readily observed structural characters. A. Jaws projecting beyond the mouth; maxillae and labium developed somewhat as in typical mandibulate insects, with conspicuous jointed palps. A median jointed tail-process. Sub-order Hetotroph. a. Body not flattened dorso-ventrally ; dorsal aspect of thorax markedly convex. Head with paired and median ocelli in addition to the large compound eyes. Abdominal segments 1-7 with exsertile vesicles, 2-9 with unjointed stylets, 10 with long jointed cerci. Family Machilidae. b. Body flattened dorso-ventrally. No ocelli; compound eyes relatively small. Abdominal segments usually without exsertile vesicles (on segments 2-7 in some Nicoletiinae), stylets usually on segments 7-9 or 8-9 only (rarely 2-9), . : ; . Family Lepismidae. B. Jaws apparently retracted within the head; mavxillae and labium highly modified with palps unjointed or absent. No median tail-process, . 2 ‘ : ; i . Sub-order Lntotrophi. ce. Cercei modified into forceps, . é 3 Family Lapygidae. d. Cerci elongate with sensory bristles, . Family Campodeidae. Family MACHILIDAE. Our knowledge of the various genera comprised in this family has been vastly extended during recent years through the work of Silvestri (04, °05, ’06, 11) and Verhoeff(?10). The latter author has deemed it advisable to recognize three distinct families instead of one. While the characters used in this discrimination—the shape and extent of the abdominal sterna and the number and arrangement of the exsertile vesicles which these bear (see Plate II, figs. 19-25, 27-31 e.v.)—are of undoubted value in facilitating classification and indicating relationship, they are not of sufticient importance to justify family-distinctions. he Machilidae, as generally understood, form such a natural and easily recognized group of Thysanura, that Silvestri is undoubtedly to be commended for followimg the older eutomologists in regarding the insects as constituting a single family. Verhoeffs divisions may eonveniently be regarded as sub-families—to be distinguished thus— A. Abdominal segments all with very small sterna, and bearing at most one pair of exsertile vesicles, é : ; Mevnertellinae. [B] 4 Proceedings of the Royal Irish Academy. B. Abdominal segments from 2nd to 7th, inclusive, with relatively large triangular sterna. a. Ouly one pair of exsertile vesicles on any abdominal segment, . F j : : : . Praemachilinae. 6. Two pairs of exsertile vesicles on abdominal segments 2-3, 2-4, 2-5, or 2-6, : : : 5 Machilinae. The last-named group (Machilinae) is alone represented in the collection from the Seychelles. Three species, all new, differ so definitely from any members of the family hitherto known that a new genus is required for their reception. Corethromachilis! geu. nov. Feelers, maxillary palps, legs, and niuth abdominal stylets scaled. Apex of mandible feebly toothed. Lacinia of maxilla with a complex “ brush ” of lanceolate bristles. Legs of second and third pairs with coxal processes ; beneath the tip of the terminal (third) segment of the foot in all three pairs a dense mass of lanceolate bristles forming a “brush” or scopula. Abdominal segments with moderately large triangular sterna, the first, fourth, tifth, sixth, and seventh with one pair of exsertile vesicles each, the second and third with two pairs each. Male with feebly jointed gonapophyses on the eighth and ninth abdominal segments; penis short, not reaching apex of the ninth sub-coxa. Female with ovipositor not, or hardly, projecting beyond ninth abdominal stylets. Type, Corethromachilis Gardineri (sp. noy.) Seychelles. This genus is of interest from the reduction to two of those abdominal segments which have two pairs of exsertile vesicles each. In the typical genus Machilis there are four segments thus provided ; in Coryphophthalmus (Verhoeff, 710) three. In this character, therefore, Corethromachilis ap- proaches Praemachilis and its allies, in which there is but one pair of exsertile vesicles on each abdominal segment from the first to the seventh, inclusive. But the most remarkable feature in Corethromachilis is found in the wonderful arrays of bristles on the lacinia of the maxilla and beneath the tip of each foot, forming the brush-like organs that have suggested the generic name. Except in the case of the aberrant C. gibba, described below, there is nothing to attract attention in the general appearance of the species. Indeed, the naturalist studying the Machilidae is struck with a monotony throughout the family in the main features of their superficial aspect, which is yet accompanied by a range of variation in details of the exoskeleton which afford reliable characters for classification. All the specimens of Corethro- 1 From xépn8pev, a broom, and Machilis. Carpenter—The Apterygota of the Seychelles. 5) machilis were collected in mountain-forest regions, offering in this respect a great contrast in habit to most Machilidae, which delight in stony places. Three species from the Seychelles are referable to this new genus; they may be distinguished by obvious characters. A. Insects of the usual Machilid build; paired ocelli transverse. a. Legs and claws short, scopulae large and very dense (fig. 17). Maxillary palps longer, Corethromachilis Gardinert. b. Legs and claws long, scopulae less dense (fig. 44). Maxillary palps shorter, . : : : C. brevipalpis. B. Metanotum raised into a great dorsal prominence (fig 45). Head with conical process between the eyes, paired ocelli shortly ovoid (figs. 46,47), . ; : é : 2 : C. yibba. Corethromachilis Gardineri sp. nov. (Plates I, II, figs. 1-26.) Paired ocelli (fig. 2 p.o.) of the usual dumb-bell shape, about a transverse diameter apart. Feelers more than twice as long as body (fig. 1); basal segment (fig. 2) three times as long as broad; regions of flagellum (fig. 3) with 16-18 segments each. Mandible (fig. 4), maxillula (fig. 5 M27), tongue (fig. 5 hy), galea of lacinia (fig. 67), and labium (fig. 12) exceptionally broad in proportion to their length. Maxillary palp one-third length of body. Legs short, claws remarkably short, and scopulae very dense (fig. 17). Stylet on second abdominal segment (fig. 20) with numerous bristles but without terminal spine. Median tail-process nearly twice as long as body ; cerci half as long as body (fig. 1). Length of body 14mm. Colour (with scaling), dark metallic purple with white rings on feelers, cerci, and tail-process. Localities—Mahé: Forét Noire district and Cascade, 1000 feet and over (August, 1905, September and Octoher, 1908); Montagne Alphonse, Cascade, 1800 feet (December, 1905). Silhouette ; forest near Mare aux Cochons, over 1000 feet (August und September, 1908). Praslin: Cétes d’Or Jungie (November, 1908). Specimens numerous in all these localities. Mr. Scott records that they all come from mountain forests, mostly among the dead leaves, both fallen and still hanging, of palms and other trees. This species is remarkable for the lateral extension of the jaws and tongue The base of the mandible (fig. 4) has a strong prominence on its outer border, while the tongue (fig. 5 hy) and the maxillulae (fig. 5 Mel) are 6 Proceedings of the Royal Irish Academi. unusually broad in proportion to their length. They show, in the main, the arrangement of parts characteristic of the Machilidae. In the mazdla, the galea (figs. 69, 9, 10) is remarkably broad and flattened, its free border merging into a delicate membranous ridge, strengthened by rib-like thickenings, and its outer corner bearing a group of minute sensory spines projecting from papillae (figs. 10,11). The lacinia (figs. 6/, 7, 8, 9) exhibits in perfection the large and wonderfully formed head to which reference has been made in the generic description. From the terminal teeth (¢) a lamella (Ja) extends on either aspect, embracing a considerable cavity from whose recesses spring more than fifty lanceolate bristles forming the characteristic “brush.’’ Such a brush, though in a far less highly developed condition, has been figured by Borner (08, Pl VI. fig. 11) from a Japanese species of Machilis, and by the present writer (715, Pl. II. fig. 64) in Petrobius. Borner calls the structure “der Mittelanhang.” The hinder edge of the lamella (see figs. 7, 8 /a) is produced into three prominent teeth (¢’) ; its front edge (fig. 9/a) has a rather sinuate margin, but ends in a single, sharp, delicate tooth (¢”)- The maxillary palp (fig. 6 p) is of the usual form, its first segment with a strong backwardly directed conical process, its succeeding six segments with the proportionate lengths 10: 10: 8:11:9:13. The whole palp measures 5 mm. in length; the terminal segment as usual bears many strong spines. The labium (fig. 12) has a relatively broad and short sub-mentum; the terminal segment of the labial palp carries a number of rows of flattened tapering sensory spines (fig. 13). The short claws and the dense scopulae below the terminal segment of each foot (fig. 17) give the insect a very characteristic appearance. ‘lhe bristles of this scopula have a regular lanceolate shape towards the tip (fig. 18), but this specialized condition can be traced through a series of gradations from the ordinary bristles of the leg. The coxal process of the second leg (fig. 15) is narrower than that of the hind-leg (tig. 16). Of the abdominal segments, the first (fig. 19) is remarkable for the reduced sternum, the second for the hairy, unspined stylets (fig. 20), and most vf the others for the sinuate sutures between the sub-coxae and the sterna. The male genital segments (figs. 24 and 25) do not call for special remark ; the penis (p) and yonapophyses are short; the latter show imperfect jointing, and bear numerous spines in rows along their inner faces (fig. 26). The stylets of the eighth and ninth segments have very long spines. In the female the gonapophyses are relatively short, with sixty-four rings on those of the eighth, and an equal number on those of the ninth, segment. Carpenter— The Apterygota of the Seychelles. w Corethromachilis brevipalpis sp. nov. (Plates IT, III, figs. 27-44; Plate V, figs. 63-6.) Paired ocelli (fig. 35 p.o.) long and narrow, almost in contact centrally. Feelers longer than the body, basal segment two and a half times as long as broad (fig. 35); regions of flagellum (fig. 36) with 14-16 segments each. Jaws of typical machilid form, “brush” of maxilla less prominent than in C. Gardinert. Maxillary palp one-fifth length of body. Legs (fig. 45) moderately long; claws long and scopula scanty (fig. 44). Stylet on second abdominal segment (fig. 28sf) with short terminal spines. Median tail-process longer than body ; cerci as long as body. : Length (without appendages) 15 mm. Colour of scaling rather paler than in C. Gardineri. Localities, —Mahé: in the mountain forests (August and September, 1908). Silhouette: forest near Mare aux Cochons, 1000 feet (September, 1908). A number of specimens from each locality, but the species is evidently less abundant than C. Gardinerc. C. brevipalpis is a somewhat larger species than C. Gardineri, but the maxillary palp (fig. 39 p) is absolutely shorter and feebler than in the latter (see fig. 6). ‘The jaws of C. brevipalpis differ less markedly from those of typical Machilids than do the corresponding structures in C. Gardiner, the mandible (fig. 37) and maxillula (fig. 38 Mal.) being of the proportions usual in the family, and the galea (fig. 39, 407) of the marila being longer than broad. The lacinia (fig. 40) is furnished with a “brush,” but its bristles are less numerous and prominent than those of C. Gardineri. ‘The first segment of the maxillary palp has its process sub-cylindrical; the proportions of the ‘other six segments are as 5:4:5:7:5:6. The legs are of the same general build as those of C. Gardineri, but longer, aud the conspicuous claws, with a slight tendency to indentation along the inner edge, and the scanty scopulae (fig. 44) make discrimination between the two species easy. ‘The abdominal segments and their appendages (figs. 27-34) correspond closely with those of C. Gardineri, except that the stylets of the second (fig. 28 st) have distinct though short terminal spines. ‘The tip of the ovipositor reaches only to the base of the spine on the ninth abdominal stylet. Its gonapophyses have from fifty-five to sixty segments each (figs. 63-64). The ovipositor in these insects is well worthy of study, though of less value than the male reproductive processes in specific determination. There are two pairs of gonapophyses on the eighth and ninth segments respectively (see figs. 63, 64 go); the bases of these are connected with the inner anterior 8 Proceedings of the Royal Irish Academy. corner of the sub-coxae to which they belong. Each gonapophysis is worked by appropriate muscles—an adductor (fig. 65 ad) the insertion of which is carried but a short distance along the appendage ; and an extensor (fig. 65 er¢), some of whose fibres are inserted close to the base, while a few, drawn out toa great length, extend right along the outer margin of the gonapophysis, almost to its tip (fig. 65 ext’). The segmentation of these appendages is well marked ; the proximal segments (fig. 65.4) bear few, feeble, and short bristles; but beyond the extremity of the ninth abdominal sub-coxae these bristles become, almost suddenly, long, stiff, and prominent (fig. 65 B), this character persisting almost to the extreme tip (fig. 65c). Hach sub-coxa of the ninth segment articulates with a small, sub-triangular, basal sclerite which from its position might be regarded as an episternum (fig. 64epst). From it originates an abductor muscle (fig. 64 abd), whose fibres pass, diverging slightly towards the axis of the body, and are inserted into the inner edge of the base of the snub-coxa. The scaling of C. brevipalpis—as of the other species of Corethromachilis —resembles that found generally in the family. ‘Two typical forms of scale from abdominal sub-coxae are figured (fig. 66), one being of a moderately broad, and the other of a narrower and elongate, type. Some of the smaller and more delicate scales are broader than long. Corethromachilis gibba sp. nov. (Plate IV, figs. 45-62.) Paired ocelli (tig. 46, 47 p.o.), short and ovoid, situated on either side of a conical prominence in front of the eyes. Jaws transversely extended, and brush of maxillary lacinia complex, as in @. Gurdineri. Maxillary palp nearly half as long as body. Thorax with the mentanotum produced dorsal- wards into a prominent hump; margin of mesonotum broadly convex laterally, and sinuate in front (fig. 45). Legs short; foot-claws short and scopulae dense (fig. 55). Feelers and tail-process half as long again as body. Cerci two-thirds length of body. Length 10mm. Colour of scaling, dark. Localities —Mahé : Mare aux Cochons, 1500 feet, in dead leaves (January, 1909, one female). Silhouette: 1500 feet, in high damp forest, among fallen, rotten palm-leaf bases, and other dead leaves in damp and shady jungle (eight specimens of both sexes, August, 1908, collected by Mr. H. Scott, who states that they jump vigorously). This insect, with the great hump on its metathorax and the outstanding conical process in front of the head, may be distinguished at a glance from all Carpenrer— Ve Apterygota of the Seychelles. 9 known Machilidae. So conspicuous are these distinctive features that the establishment of a distinct genus for the species might be thought desirable by some entomologists. But in the structure of its jaws, feet, and abdominal segments and appendages the insect resembles so closely the two Corethro- machilis already described, that it seems reasonable to consider it cogeneric with them. It is suggestive that in C. gibba the tendency to develop conical out-growths should be displayed both on the head and the metathorax. In the feeler the basal segment is three times as long as broad, and the regions of the flagellum have about 18-20 segments each (fig. 48). The arrangement of the ocel/z is most remarkable, the median one looking directly downwards and those of the pair being placed close together on either side of the conspicuous prominence in front of the eyes (figs. 46, 47). The eyes are in contact for a comparatively short distance along the median axis of the head. Asin @. Gardineri, the mandible (fig. 49), tongue (fig. 50 hy), maxillula (fig. 53 Ml), marillary galea (tig.51 9), and labiwm (fig. 52) are exceedingly broad in proportion to their length. The maxillary lacinia has a head with complex brush, resembling that of C. Gardineri (figs. 7, 8, 9) so closely that it is needless to figure the details. The /egs (figs. 54, 55) also are very like those of C. Gardineri; the scopula, however, in C. gibba is rather smaller, and the claws are a little longer than in the former. ‘lhe abdomen of C. gibba is relatively small as compared with that of the other species, as may be seen by comparing the outlines of the abdominal segments (figs. 56-60) with those of the corresponding structures on Plate II. On the second segment (fig. 56) the stylet has a short but distinct spine. The ninth stylet (fig. 60 s¢) has a slender, acute spine almost its own length. The male gonapophyses of C. gibba are weakly developed and feebly jointed (figs. 61-62). It is well known that various interpretations of the genital armature of insects have been given by different students. The term yonapophyses emphasizes the correspondence of these structures in the Thysanura with those in the Orthoptera and other insects which have a typically developed male and female armature. Terms such as “telepodite” and “ parameron ” have been applied by some authors who, like Escherich (04, pp. 23-6), regard a genital process as comparable to the terminal portion of a thoracic leg. Silvestri (05, pp. 794-7) has argued convincingly in favour of the opposite view: that the abdominal stylets, rather than the gonapophyses, are to ve regarded as appendicular; and he has brought forward some reasons for con- sidering the latter as homologous with the exsertile vesicles on the unmodifiea abdominal segments. It is at least suggestive that the genital segments of the abdomen never bear exsertile vesicies in the Thysanura. R,I.A, PROC., VOL. XXXIII., SECT. B. [C] 10 Proceedings of the Royal Irish Academy. Family LEPISMIDAE. This family is fairly well represented in the fauna of the Seychelles, as shown by the collection now described, and there can be no doubt that further species await discovery. Students of the Lepismidae owe much to Escherich, whose beautiful monograph (’04) stands as a foundation for modern systematic work. He divides the family into three sub-families :— A. Inner edge of maxillary lacinia with teeth and bristles. Head never longer than prothorax. a. Sub-coxae of genital segments broad and flat, covering bases of gonapophyses. Eyes present. Terminal seg- ment of maxillary palp without sensory papillae. Body always scaled, . - - : . Lepisminae. b. Sub-coxae of genital segments narrow, not covering bases of gonapophyses. Eyes wanting. Terminal segment of maxillary palp with conspicuous sensory papillae. Body scaled or unscaled, j : 3 : . Nicoletiinae. B. Inner edge of maxillary lacinia smooth. Head longer than pro- thorax. Eyes present. Body unscaled, . : . Maindroniinae. The first and second of these sub-families are represented in the Seychelles; the last is known to include only a single species from Arabia. It is noteworthy that all the Seychelles Lepismidae belong to well-known and widespread genera, in contrast to the Machilidae. which are represented in the archipelago by a distinct and peculiar genus. « LEPISMINAE. The Lepisminae in the collection are distributed among four genera— Lepisma, Isolepisma, Ctenolepisma, and Acrotelsa, which are thus distin- guished :-— A. Bristles on head and body-segments simple. Tenth abdominal tergite rounded, truncate, or emarginate. a. Bristles on face and terga arranged singly, not in “combs,” : : 5B 2 : : Lepisma. b. Bristles on face and terga arranged in “ combs,” Isolepisma. B. Bristles feathered. c. Tenth abdominal tergite long and acutely pointed, Acrotelsa. d. Tenth abdominal tergite truncate or emarginate, Ctenolepisma Carpentrer— The Apterygota of the Scychelles. 11 In distinguishing genera, Escherich lays great stress on the nature and arrangement of the bristles, which are unfortunately often knocked off in preserved specimens. The scars marking their points of insertion are, how- ever, usually conspicuous. Lepisma Linné. This, the best-known genus of the family, has the little household European “Silver-fish,” Zepisma saccharina Linn., as its type species. In the collection from the Seychelles the genus is represented by a single myrmecophilous species, which is apparently new. A number of Lepismae from various regions are well known as guests of ants and termites. Lepisma intermedia sp. nov. (Plate V, figs. 67-70). Thorax moderately convex laterally, abdomen evenly narrowed behind ; body about three times as long as broad. Feeler half as long as body. Eyes small and round. Abdominal terga with two pairs of dorsal bristles. Tenth abdominal tergum twice as long as ninth, sinuately emarginate behind. median tail process three times, cerci twice as long as tenth tergum (fig. 67). Ninth sub-coxa with inner process only slightly longer than outer (fig. 68). Length 3-4 mm. Colour of scaling deep brownish violet dorsally, ventral surface, feelers, legs, and appendages generally pale yellow. Localities.—Mahé: |.ong Island, from a nest, in decayed log, of Pheidole punctulata, an ant known in both Africa and Madagascar (July, 1908, three specimens). Félicité (1908, two specimens). Unfortunately all the specimens of this little Lepisma are dry and carded, so that it is not possible to make out many structural details. The legs have rather wide shins (fig. 69), and feet with the first and third segments each slightly longer than the second. The shin bears at its outer tip the broad spur commonly found in this family, and some flattened sensory bristles (fig. 70) feebly hooked at the tip. JZ. intermedia comes nearest to L. Braunsi Escherich (from South Africa), and LZ. indica Mscherich (04, pp. 50-51), ditfering by its smaller size, relatively longer median tail-process, and shorter inner ninth sub-coxal processes. It resembles LZ. Brawnsi in its emarginate tenth tergum, and Z. indica in its round eyes. Isolepisma Escherich. This genus was established by Escherich (04, pp. 61-2) for a single wide-ranging tropical species in the description of which no clear indication is given of what are considered generic as contrasted with specific characters. Now that a second species has to be described, it is possible to give a more precise diagnosis of the genus. [C2] 12 Proceedings of the Royal Irish Academy. Body somewhat narrow; thorax not markedly broader than abdomen, whose tenth tergum is distinctly longer than the ninth. Bristles, simple or bifid at the tip, arranged in tufts on the head, and in “combs” of two or three each on the thoracic and abdominal terga. Two or three pairs of abdominal stylets. Isolepisma bisetosa sp. nov. (Plate VI, figs. 71-82.) Length 7°5 mm. Feelers, median tail-process and cerei shorter than the body (fig. 71). Terga of thoracic segments with a “comb” of two strong bristles at the hinder edge on each side of the middle line, and five or six lateral “combs” (fig. 71). Terga of abdominal segments (II-VIII, inclusive) with two “combs,” usually of three, but occasionally of four bristles at the hinder edge on each side, and a comb of two bristles nearer the median line (fig. 71). Process of sub-coxa of ninth abdominal segment in -both sexes acuminate, nearly half as long as the stylet (figs. 80, 81). Eighth abdominal segment of male (fig. 80), and seventh and eighth of female (fig. 81), with short stylets. Ovipositor in female (fig. 82) elongate, projecting beyond the tip of the ninth stylet for at least the length of the latter. The above features serve to distinguish this species from J. ¢risetosa, Esch. (04, pp. 62-3, Pl. I, fig. 1), which is rather smaller, has the feelers and caudal process as long as the body, and possesses paired combs on the thoracic segments consisting of three bristles each. Localities.—Seychelles: Bird Island (1908); Ile aux Récifs (H. P. Thomasset); Long Island (July, 1908). Aldabra (J. C. F. Fryer, 1908-9). Farquhar (30th September, 1908). Providence: Cerf Isl. (J. 5. Gardiner, 3rd September, 1905). Amirante, Eagle Island (J. 8. Gardiner, May- September, 1905). This species is apparently abundant, as it is represented by numerous examples from most of the localities. It is very closely allied to /. trisctosa, Escherich—the only species of the genus hitherto known—which has apparently a cireumtropical range (Brazil, West Africa, Malay Archipelago). It may probably be regarded as having been differentiated from J. trisetosa during a long period of isolation. Sufficient material of this species is available for an examination of the principal structures of the head. The labrum (fig. 72 /br) is a short, broad plate, with arched front edge, hinged on to the face (fig. 72 /), which bears numerous slender bristles, and two series—three on either side—of broad, bifid ones, like those that project in front of the clypeus. The mandible differs distinctly in shape from that of Ctenolepisma, figured by Escherich CaRPENTER— The Apterygota of the Seychelles. 13 (04, Pl. II, f. 26). There is a rounded condyle at the base, and the inner hind edge projects around the hollow into which the fibres of the adductor muscles pass (fig. 75). On the outer edge of the mandible are a number of bristles, most of them bifid like those of the head, but a few simple; there are also some bifid bristles along the front inner edge, near the molar area, which is beset with a number of short, strong spines (fig. 76). Beyond these is a blunt projection, and the apex has three prominent but not very sharp teeth ; a slight difference between the apices of the right and left mandibles is noticeable. The mandible is worked by at least six distinet muscles. There are two retractors (fig. 75 re), inserted into the inner median ridge— one by a single tendon, and the other by a number of slender tendons. A posterior adductor muscle (fig. 75 ad. p.), with a tendinous insertion, pulls at the basal region of the mandible, while three median adductors (fig. 75 ad), one large and two smaller, with fibres radiating so as to be inserted along the inside of the outer wall, serve to draw the mandible strongly towards its fellow. A closely similar arrangement in Lepisma saccharina has been described and figured by Borner (09, pp. 104-5, fig. 2). The tongue in Isolepisma (fig. 77 hy) is small, with the tip hairy, and the peduncles almost parallel. The mawtllulae (fig. 77 imal) have a roughened apex, with fine, short bristles. No differentiation into galea and lacinia—so apparent in a Machilid maxillula—is here distinguishable. Hansen twenty years ago pointed out that the maxillulae of the Lepismidae are feebly developed as compared with those of the other Thysanura, recalling the con- dition found in earwigs (93), but no entomologist has hitherto given attention to these interesting structures in this family. The mavilla (fig. 78) resembles in its main features that of a Lepisma figured by Escherich (04; Pl. II, fig. 29). The tip of the lacinia has two strong teeth; its inner edge is drawn out into seven delicate teeth, forming a “comb” (fig. 787), proximal to which are eight prominent bristles. The musculature of the maxilla is like that found in biting insects generally ; there is a strong protractor muscle (pr.) on the inner edge of the cardo (c), while the muscles for working the galea and iacinia (g. m. and /. m.) originate at the proximal end of the stipes, and have their fibres converging to the insertions at the bases of the lobes. The /abewm (fig. 79) has a short and broadly arched sub-mentum (s. m.), the mentum (m), galeae (g), and laciniae (/) being small, somewhat rugose and spiny ; the distal edge of the lacinia is drawn out into a delicate ridge. The four-segmented labial palp has a very broad and blunt terminal segment (fig. 79 p). In both sexes the inner process of the ninth abdominal sub-coxa (figs. 80, 14 Proceedings of the Royal Irish Academy. 82s.¢., 81) is elongate, acute, and spiny. The penis (fig. 80 pe), as usual in Lepisma and allied genera, is short; the ovipositor projects far beyond the tip of the ninth abdominal stylets (fig. 82). Ctenolepisma Escherich. This genus was founded by Escherich ('04, p. 75) te include Lepismidae characterized by the possession of numerous “ combs,” of feathered bristles on the thorax and abdomen, four being present on the greater number of the abdominal terga, and by the rounded truncate or emarginate edge of the tenth abdominal tergum. It includes a number of species, whose collective range extends over the tropics and warmer temperate regions of the globe. Ctenolepisma longicaudata Escherich. A single female from lélicité, Seychelles, 1908, is evidently referable to this species, which Escherich described (04, pp. 83-4, fig. 51) from specimens found in houses in South Africa, and to which he referred doubtfully a specimen from Guinea. ‘The presence of the insect on the Seychelles confirms his suggestion that it would be found widely distributed in the Ethiopian Region. Acrotelsa Escherich. This genus was established (’04, p. 105) for some Lepismids of relatively large size, distinguished from allied genera by the tenth abdominal tergum being long and pointed. Escherich includes this genus in the sub-family of the Lepismatinae, among the diagnostic features of which (op. cit., p. 36) he mentions the absence of sensory papillae on the terminal segments of both maxillary and labial palps, such sensory papillae being present in the Nicoletiinae. In Isolepisma, and probably in most of gthe species of Lepisminae, both maxillary and labial palps are without these papillae, and so are the maxillary palps in Acrotelsa. But in the three species of Acrotelsa from the Seychelles collection the labial palps have very con- spicuous papillae on the terminal segment, and one of these species is clearly identical with that described by Escherich as Acrotelsa collaris (Fab.). Silvestri also has described some species of Lepsima with similar structures (13, pp. 8-11). It is necessary, therefore, to revise the diagnostic characters of the Lepisminae, and to recognise that in this feature members of the sub- tamily may approach the Nicoletiinae. The sensory papillae of Acrotelsa are shown in figs. 88, 89 (A. elongata, sp. nov.), 98, 99 (A. Scotti, sp. nov.), and 101 (A. collavis). In the two former species the five papillae are arranged in a single row along the broad end of the terminal segment, while in A. collaris CarpEnreR—The Apterygota of the Seychelles. 15 they form a proximal row of three and a distal row of two. The cuticle of the papilla has a roughened, wrinkled surface, and is protrusible from a sub-cylindrical projection of the general firm cuticle of the appendage. In the examination and delineation of the insects of this genus I grate- fully acknowledge some valuable help from my colleague, Miss A. J. Reilly, A.R.C.Sc. The three species of Acrotelsa from the Seychelles may be distinguished thus :— A. Large, indoor species, over 15mm. No dorsal combs of bristles on thoracic terga, . : : A. collaris (Fab.). B, Small species under 12mm. A pair of dorsal “combs ” on each thoracic tergum. a. Inner process of ninth abdominal sub-coxa very long, A, elongata, sp. nov. b. Inner process of ninth abdominal sub-coxa of normal length, : 0 5 : : . A, Scotti, sp. nov. Acrotelsa collaris (Fab.). This widely distributed species—the common house-lepismid of the tropics—has already been recorded from the Seychelles (Escherich, ’04, p. 108). It is found in both hemispheres, and there can be little doubt that its wide range is largely due to commercial importation. The presence of sensory papillae on the terminal segment of the labial palp has already been mentioned. ‘The insect has been well figured by Oudemans (790, Pl. VI, fig. 1) and Escherich (704, Pl. I, fig. 3). Localities.—Seychelles: Bird Island (1 male, 1 female); Mahé: Port Victoria (3 specimens, dried and carded, 1908-9), Round Island (1 specimen, dried, July, 1908). Coetivy Island (3 specimens, dried and carded, 1905). Aldabra: Picard Island (1 male, 1 female, January, 1909). Acrotelsa elongata sp. nov. (Plate VII, figs. 85-90.) Body-form elongate, narrow. Terminal segment of labial palp (figs. 88, 89) broad and sub-globose with five.sensory papillae. Each thoracic tergum with a dorsal and ten lateral “combs” of bristles on each side. Abdominal terga ii-vill with a dorsal and a marginal “comb” on each side. Tenth abdominal tergum acuminate, longer than broad, with four marginal “combs” on each side (fig. 83). Inner processes of ninth abdominal sub-coxa very elongate, their tips almost reaching the extremity of the ovipositor (fig. 90). 16 Proceedings of the Royal Irish Academy. The feelers, cerci, and tail-process are unfortunately very imperfect in all the specimens. Length, 9mm. Colour, pale with brown scalding; feelers, cerci. and tail- process dark-ringed. Localities, —Aldabra ( 1908, coll. J. C. F. Fryer, five females). This species is very closely allied to the North Australian A. producta Escherich (04, pp. 111, 112) with which it agrees in the immensely elongate ninth sub-coxal processes. A. producta, however, has a much narrower terminal segment to the labial palp, and a tenth abdominal tergum that is shorter than broad (Escherich 1. c., fig. 45). It is remarkable that this most abnormal form from Aldabra should have its nearest ally on the Australian continent. Some notes on the jaws of this Acrotelsa may be of interest. The mandible (figs. $4, 85) is relatively longer and narrower than in Isolepisma (see fig. 75), and more convex externally than in Ctenolepisma (Escherich, 1904, Plate I, fig. 26). In the group of fine spines at the molar area and the bifid bristles just proximal thereto this mandible agrees with those of Lepisminae generally. The marilla (fig. 86) calls for little remark; the somewhat acuminate tip of the galea projects well beyond the lacinia, which has the inner edge, just proximate to the apical tooth, serrate, with seven delicate lanceolate “comb” teeth, and armed with six prominent bristles (fig. 87). The labium (fig. 88) has the sub-mentum, mentum, and lobes very broad, as well as the terminal segment of the palp, along the edge of which are five oval sensory papillae in a row (fig. 89 s. p.). Acrotelsa Scotti sp. nov. (Plate VIII, figs. 91-100.) Body-form elongate, narrow. Terminal segment of labial palp (fig. 98 p), broad and sub-globose, arranged with five sensory papillae (fig. 99 s. p.). Each thoracic tergum with a dorsal and twelve lateral “combs ” of bristles on each side. Abdominal terga ii-viii with a dorsal and a marginal “comb” on each side. Tenth abdominal tergum acuminate, longer than broad, with four marginal “combs ** on each side (fig. 91). Inner process of ninth abdominal sub-coxa moderately long, surpassing the tip of the relatively short ovipositor. Gonapophyses slender, cylindrical, feebly segmented (fig. 95). As with the previous species, the few specimens are all very imperfect as regards feelers, cerci, and tail-process. Length, 11 mm. Colour, pale with brown scaling. Localities —Aldabra: Ile Esprit and Takamaka (November, 1908, four females, J. C. F. Fryer, coll.). Carpunrer—The Apierygotu of the Seychelles. 17 This species is not closely allied to any mentioned in Escherich’s “System” (1904), but it comes very near to A. Voeltzhowi, subsequently described by him from Madagascar (10), which differs from A. Scotti mainly in having only seven marginal combs on the thoracic terga and only three on each side of the tenth abdominal tergum. The jaws of A. Scotti resemble rather closely those of A. elongata. Some details for comparison are shown on Plate VIII (figs. 92-3, 96, 98), but they do not call for special description. In the leg (fig. 100) the shin has the usual prominent spur overhanging the base of the proximal segment of the foot, whose third (distal) segment carries between the two claws a slender, almost straight, claw- like empodium. NICOLETIINAE. Escherich (04) included four genera—Atelura, Lepidospora, Nicoletia, and Trinemophora in this sub-family ; the two former only are represented in the collection from the Seychelles. Atelura, as understood by Escherich, includes a number of small, scaled, blind insects, resembling Lepismae in general aspect, which live as the guests of ants and termites. Silvestri (’08) has referred some of Hscherich’s species of Atelura to several distinct genera. Only a few species of Lepidospora are known from various tropical and sub- tropical countries ; these are large, scaled, free-living, bristle-tails. Possibly some examples of the wide-spread scaleless Nicoletiae await discovery in the Seychelles. Atelura Heyden. Two dried specimens represent this genus in the collection. They appear identical with or very close to one of Escherich’s species of this genus in the wide sense, and as they are unsuitable for microscopic examination, I refrain from any attempt to discuss their relationship among Silvestri’s groups. Atelura nana Escherich. In the small size (under 2 mm. long), pale colour, and long dense bristly covering (two or three rows on the thoracic segments) the Seychelles speci- mens agree closely with this species described from South Africa (’04, p. 127, fig. 53), where it was found in nests of Pheidole punctulata Mayr. Locality—Mahé: Round Island (from nest of Pheidole in broken stone, 19th July, 1908, two specimens). Lepidospora Hscherich. Escherich founded this genus ('04, pp. 131-2) for the reception of the species L. Braueri, described by him froma single male specimen which was R-I.A. PROC., VOL. XXXIII., SECT. B. [D] 18 Proceedings of the Royal Irish Academy. brought from the Seychelles to the Hamburg Museum. He referred to this genus another species L. gracilis (J.c., p. 178) founded on a single female from Sumatra. Subsequently Silvestri (08a, pp. 382-4) described LZ. Hscherichu from Corfu, ZL. ceylonica (10, pp. 95-6) from Ceylon, and L. afra (088, p. 12) and ZL. meridionalis (13, pp. 12-13) from South Africa. These species all agree in their fairly large size and narrow build, in which they resemble the Nicoletiae,while, unlike the members of that genus, their bodies are scaled. The species of Lepidospora, thus apparently few in number and scattered in their distribution, show a strange divergence in the form of the ovipositor. In LZ. Braueri (whose female is described below) the gonapophyses (Plate X, figs. 117, 118, go. 119) are thick, unjointed, and spindle-shaped, like those of Atelura; the same form of ovipositor is found also in LZ. Escherichti Silvestri. On the other hand, in Z. gracilis Escherich and L. meridionalis Silvestri the gonapophyses are elongate, slender, and jointed like those of Nicoletia and of most of the Lepisminae. In Z. ceylonica Silvestri they are thick and jointed. In this character, therefore, the structure of Lepidospora indicates an interesting transitional condition. A primitive feature of Lepidospora is seen in the presence of eight pairs of abdominal stylets; this large number, characteristic of the Machilidae, is reduced to two or three pairs in most species of the Lepismidae. Lepidospora Braueri Esch. (1904). (Plates IX, X, figs. 102-120). This very curious and interesting species is described and figured in Escherich’s Monograph from a single male, collected in the Seychelles, and preserved in the Hamburg Museum. The present collection contains several specimens of both sexes (all from the high level forest regions), so that some account of the structural features of the insect can be given. Attention has been especially paid to the jaws and to the ovipositor in the females. Localities.—Mahé, in mountain forest: Cascade, 2,000 feet (one male, one female, 3rd December, 1905); Montagne Alphonse, Cascade, 1,800 feet (one female, 4th December, 1905) ; Mare aux Cochons, 1,500 feet, in dead leaves (one female, January, 1909). Silhouette: highest point, 2,467 feet, in damp earth under dead leaves (one male and four females, several immature, 2nd September, 1908); forest near Mare aux Cochons, over 1,000 feet (9th September, 1908, one immature). A description of the jaws and ovipositor of a Mediterranean species of Lepidospora (LZ. Escherichii from Corfu) has been given by Silvestri (08 a, pp. 382-4, figs. 18,19). The mandibles of L. Braueri (Plate IX, figs. 104-105) Carpenrer—Vhe Apterygota of the Seychelles. 19 resemble those of Silvestri’s species rather closely, even to a characteristic difference between the right (fig. 104) and the left jaw (fig. 105) in the form and arrangement of the teeth. The large retractor muscle of the mandible (fiz. 104 re) is broad and strap-shaped; the fibres of the principal adductor (fig. 104 ad) ave gathered into a narrow tendon, whence they radiate to the inside of the convex border of the appendage, as in the Lepismatinae. The maxilla resembles that of Nicoletia as figured by Escherich (04, Plate II, fig. 27), but the palp in Lepidospora is relatively much longer. The terminal segment of the palp carries at its tip four knob-like sensory prominences beset with numerous fine hairs (figs. 107, 110, s.p.); it also bears, a little behind the tip, a flattened, annular structure (fig. 110 s.7.), which may also be regarded as a sense-organ. At the tip of the remarkably slender galea (fig. 107 ga.) are two peg-like spines, probably sensory. The lacinia (fig. 107 /., fig. 108) has two teeth at its extremity, and carries on its inner border a most beautiful and elaborate “comb-process” (fig. 108 ¢.p.). This process is beset towards its tip with a double row of strong spines (fig. 109), while at the base there is a series of four or five complex, flattened spinose processes, one branch of each being bifid at the extremity (fig. 108). The tongue (fig. 106 hy) is relatively narrower than that of Isolepisma ; its tip is emarginate, with a small central prominence. The mazillulae (fig. 106 mai.), rather long and narrow in form with rounded extremity, are beset with oblique, parallel rows of fine hairs. The labivm (fig. 111) resembles rather closely that of Nicoletia as figured by Escherich (04, Plate II, fig. 32), the basal plate, galeae, and laciniae being relatively longer and narrower than the corresponding parts in the Lepisminae. In Nicoletia, however, the tip of the lacinia is bitid, while in Lepidospora it is simple. The very broad terminal segment of the palp bears six circular, cushion-shaped sense-organs, each beset with numerous fine hairs (fig. 111). A thoracic /ey in Lepidospora consists oi coxite, trochanter, thigh, shin, and three-segmented foot (Plate X, fig. 112). The coxite carries several bifid hairs. At the tip of the shin is a strong, claw-like spine overhanging the base of the foot. At the tip of the foot are two strong claws, beset on the basal half with fine short hairs; between the claws projects a slender empodium, bluntly rounded at its extremity (fig. 113). No feature of special interest is presented by the sterna of the abdominal segments from the second to the seventh. Each carries, as usual in the group, a pair of stylets and a pair of protrusible vesicles (fig. 114). Turning to the terminal abdominal segments and their reproductive processes, we find that those of the male have been already well figured [D 2} 20 Proceedings of the Royal Irish Academy. by Escherich (04, text fig. 58, p.151). I give now, however, a drawing (Plate X, fig. 116) of the ventral view of the tenth abdominal tergum, showing how the stout spines are arranged around the latero-posterior edges of the sclerite, and demonstrating also the conical apodemes (fig. 116 ap) and the muscles connected with the cerci and the median appendage. In the female the tenth tergum is simpler and relatively narrower (fig. 120), without the stout spines that characterize the male, but carrying a pair of long, terminal bristles. The ovipositor of Lepidospora Braueri resembles that of ZL. Hscherichii, which Silvestri has figured in outline. The large sternum of the eighth segment (Plate X, figs. 117, 118, viii) has a rounded hind margin beset with strong bristles; the sub-coxa is broad (fig. 118 s.c.8), with the stylet inserted, as usual, in a notch near its inner edge. The anterior gonapophysis or process of the ovipositor (figs. 117, 118, go. 8) is large, broadly expanded in the middle, and with a blunt tip beset with hairs and a few short spines. The sub-coxa of the ninth segment (figs. 117, 115, s.c.9), on the other hand, is narrow and elongate, constricted centrally; the sub-coxa (s.c. 9) and the stylet together are nearly as long as the gonapophysis (go. 9), which is narrower than the corresponding process of the eighth segment, shows very imperfect jointing, and carries, near the tip on the inner ventral aspect, a row of thick, curved processes, forming a comb- like structure (fig. 119). In a young individual in which no gonapophyses can yet be distinguished, the stylets of the ninth segment ave longer than those of the segments in front, and the sub-coxae are already prominent (Plate X, fig. 115). The ovipositor of this species of Lepidospora is noteworthy on account of the poorly developed jointing of the gonapophyses. Escherich has given reasons for believing that this jointing, so apparent in the Lepisminae and nature. If this view be accepted, the condition in Lepidospora Braucrt must be regarded as primitive. It has already been pointed out that in ZL. gracilis Escherich and in L. meridionalis Silvestri the gonapophyses of the female are long, with ” in Atelura, for example, is of a “false” and “secondary conspicuous jointing. Family IAPYGIDAE. ‘The members of this family are easily distinguished from other Thysanura by the transformation of the hindmost abdominal appendages into a pair of forceps —a character which gives them the appearance of tiny earwigs. They resemble, however, the Campodeidae and the Collembola in the structure of their jaws, which are for the most part retracted into the head-capsule. The typical genus Japyx was established long ago by Haliday (64) for an Italian Carpenrnr—The Apterygota of the Seychelles. 21 species J. solifugus. During recent years a large number of species have been described from various parts of the world, ranging from the Mediterranean countries and the United States to New Zealand and Chile. It is not surprising, therefore, to find the genus represented in the Seychelles collection. Tapyx Haliday. Tapyx silvestris sp. nov. (Plate XI, figs. 121-134.) Feelers 25-28 segmented. Legs somewhat slender; foot four-fifths length of shin, bearing two slightly curved claws, with feeble tooth on inner edge, anda short empodium (fig. 121). First abdominal sternum of male (fig. 126) with a small median rounded projection on the hinder edge, and a pair of vesicles, over which are two “comb ”-series of small sharp bristles. Sixth abdominal tergum evenly rounded behind; seventh with the hind corners very slightly produced ; tenth distinctly longer than broad (fig. 127). Forceps nearly as long as tenth segment; slightly asymmetrical, with the tips of the cerci strongly turned inwards, and the right internal tooth nearer the base than the left (figs. 127, 134). Length 7mm. Colour varies from whitish to chestnut in different specimens—all apparently adult. Localities.—Mahé: Forét Noire, at and over 1,000 feet elevation (5 specimens, both sexes, 1908). Silhouette (4 males, 1908). This species is remarkable for the comparatively small number of antennal segments. In Verhoeff’s synopsis of the genus (’04) no species with less than thirty segments in the feeler is mentioned. According to Silvestri (08a, p. 889) the true L. solifugus Haliday has twenty-eight, while J. anodus Silvestri (05, pp. 788-9) from Chile, has only twenty-seven. ‘The latter, however, may be readily distinguished by the relative thickness of its legs, and the absence of internal teeth on the forceps, from the Seychelles insect. From J. solifugus, the species now described differs markedly in the armature of the first abdominal sternum as well as in the comparatively broad and blunt processes at the hind corners of the seventh tergum. J. silvestris is remarkable in apparently possessing only one “auditory” bristle beneath each of the three antennal segments (fourth, fifth and sixth), which usually carry three or four such structures (fig. 124). From comparison of the specimens in this collection it appears that the feelers of Iapyx are capable of a high degree of contraction and extension. All the individuals from Mahé had the feelers presenting the appearance shown in fig. 121, while in two of the Silhouette specimens they were very 22 Proceedings of the Royal Irish Academy. much shorter and markedly thickened a little beyond the base (fig. 122). Naturally the first conclusion drawn was that the latter must belong to a distinct species, but when a third insect from Silhouette was seen to have one long and slender, and one short and thick feeler, and a fourth to have the basal half of its feeler thick, and the distal half slender (the junction between the two sections is shown in fig. 125), it was clear that these appendages must be capable of great modification in appearance. This was confirmed when one of the short and thick-feelered insects was transferred from alcohol to caustic potash with the result that the distal half of the feeler lengthened out, resuming its previous contracted condition when the specimen was passed on into glycerine. Study of well-cleared specimens show that each segment of the feeler consists of a cup-shaped middle region broadening distally, covered with firm cuticle, while the proximal and terminal regions are covered with thin flexible cuticle, which has a wrinkled surface in partly contracted specimens. What happens on contraction is that these firm, cup-shaped regions are pulled back into each other, the flexible intermediate tracts being invaginated (fig. 123). For this purpose the feeler is provided with two strands of longitudinal muscle. No reference to this interesting change of appearance seems to have been made hitherto, and it will be necessary for systematists to consider it in future when describing the feelers of insects of this family. The jaws of the South European Japyx have been well described by Meinert ('65), von Stummer-Traunfels (91), and Borner (’08), and as those of the Seychelles species resemble these very closely, it is needless to dwell upon them, though considerable difference of opinion has been expressed as to the homology of the structures usually regarded as maxillary. The under surface of the head of J. silvestris (fig. 125) shows the features usual in the labium of this family, with the stumpy, bristly, unjointed palps (p) that characterize the typical genus Iapyx. Very little attention seems to have been paid to the genital armature in Iapyx. Grassi (88, pp. 569, 572, pl. iv, fig. 47, pl. v, fig. 52) described and figured somewhat diagrammatically the external reproductive organs in both sexes, and Verhoeti drew the male (’04, pl. v, fig. 22) and female (703, pl. xviii, fig. 8a) structures of Heterovapyr novae-zeelandiae. In both sexes there is a small sub-semicircular chitinous plate connected by tlexible cuticle with the hind edge of the eighth abdominal sternum, behind which it is usually reflected. When protruded, therefore, it appears between the eighth and ninth sterna. In the male (figs. 132, 133) this plate has a marginal row of long bristles, and its ventral edge is beset thickly with short spines. Ventral to this plate extends a straight, hairy ridge (fig. 132 7), from CarPENTtTER— The Apterygota of the Seychelles. 23 the two ends of which project the short, unjointed, bristly gonapophyses (go.). The crescentic opening of the ejaculatory duct is just hidden by the above- mentioned ridge when the organs are seen ventrally. In fig. 132 is shown the shape of the opening, as seen throuch the thin, translucent cuticle of the genital plate viewed from the dorsal aspect. The genital plate in the female (fig. 130) has the same form as in the male, but its ventral surface is feebly granulated, bears no spines, and carries only a few long bristles. The vulvar opening is between this plate and a transverse ridge which projects dorsal to it; between the two lobes (fig. 130 7) which bound this flap the central spermathecal opening (fig. 130 spc) appears, while external to the lobes are the gonapophyses (fig. 130 go.), less prominent than those of the male. The lobes and the gonapophyses are evidently the inner and outer “papillae” as figured by Grassi; their arrangement in Heteroiapyx novae-zeelandiae, as sketched by Verhoeff (’03, pl. xviii, fig. 8a), is very similar. At the extreme tip of the female gonapo- physis is a bluntly conical papilla, around the apex of which five or six minute bristles form a ring (fig. 131). Family CAMPODEIDAE. The interesting little insects comprised in this family agree with the Tapygidae in their retracted jaws, but differ in the nature of the hindmost abdominal appendages, which are elongate tail-feelers or cerci, as in Thysanura generally. Very little is as yet known of tropical Campodeidae ; being blind insects living in soil and such concealed surroundings, they are seldom collected, and, being very fragile, imperfect and unrecognizable specimens are commoner than those fit for description. The Campodeidae from the Seychelles are few in number but highly interesting, as the species represented clearly belongs to the little-known genus Lepidocampa (Oudemans, 1890), whose members are distinguished from all other Campodeidae by being partially clothed with scales. Lepidocampa Oudemans. This genus was established by Oudemans (90, pp. 76-7) for an Indo- Malayan species Z, Weberi, inhabiting Sumatra, Java, and Flores. Silvestri (99) found what he regarded as this identical species in Argentina, and afterwards (01, p. 242; 05, p. 777) mentioned its presence in other parts of South America—Brazil, Paraguay, and Ecuador. Oudemans gives the number of antennal segments (over thirty), as he observed it in the Malayan insects, as a generic character, but Silvestri states that in the Argentine specimens the number of segments in the feelers varies from 22 to 32, 24 Proceedings of the Royal Irish Academy. Just 22 ave present in the few Seychelles specimens that possess a perfect feeler or two, so that, as far as this character is concerned, they might be referred to Oudemans’ species, which they evidently resemble closely in size and general appearance. In the minute structure of the jaws, legs, and abdominal appendages, they appear to differ distinctly from the figures which Silvestri (99, pls. 6,7) has given of the South American Lepidocampa ; and specific identity in wingless insects from such widely separated areas is unlikely. Indeed, Silvestri (01, p. 242) admits that comparison of types would be necessary to verify the reference of his Lepidocampa to Oudemans’ species. For the present, therefore, it seems advisable to describe- the Seychelles specimens under a new name. Lepidocampa fimbriatipes sp. nov. (Plates XII, XIII, figs. 135-157.) Feeler with twenty-two segments, four “auditory” bristles on segments 3-6, inclusive, and a rosette-shaped antennal organ at tip of terminal segment (figs. 155-8). Mandible with four prominent apical teeth, and a lacinial “comb” of eleven teeth (figs. 141-2). Legs asin Z. Weberi Oudemans: the laminate pulvilli bearing on each side a series of stiff, bristly outgrowths (fig. 146). Stylets of first abdominal segment in male (fig. 153) broadened distally with numerous spines, in female bluntly tapering, with a few terminal papillae (fig. 147). Stylets of abdominal segments 2-7 each with two con- spicuous stout terminal spines (fig. 148), Telson and anal valves acutely pointed (figs. 149-150). (Cerci wanting in all specimens.) Length 3°5 mm. Colour, rich brown—the scales showing a golden lustre when dry. Localities.—Silhouette. Mahé, Forét Noire, 1000 ft. A few specimens only from each island. The rarity and interest of this genus make it advisable to enter into some details with regard to the structure of our Seychelles species for comparison with those given by Oudemans and Silvestri for the Malayan and Neotropical forms respectively. The Campodeoid aspect, combined with the restricted clothing of most characteristically shaped scales (fig. 140) —present on thorax and abdomen, but not on head or legs—makes a Lepidocampa easily recog- nisable, The jeelers are shown by the Seychelles specimens to be capable of retraction and extension like those of Iapyx mentioned above (pp. 21-2). Silvestri saw and figured (99, pl. 7, fig. 10) the wntennal organ at the tip of the feeler’s terminal segment; as seen from the side, it appears—as shown in his drawing—as a relatively large papilla surrounded by bristles. In Carpenter— The Apterygota of the Seychelles. 26 L. fimbriatipes (fig. 138) it is seen in surface view to consist of four prominences arranged rosette-wise, and surrounded by a cuticular thickening, close to which project a number of stout, thick, elongate spines. Silvestri overlooked the “auditory” bristles on the antennal segments from the third to the sixth, inclusive (figs. 136, 137). They agree closely with the similar structures found in Campodea ; each bristle springs from the centre of a conspicuous cup-shaped depression of the cuticle. There are two of these sensory bristles on the lower, and two on the upper, aspect of each antennal segment bearing them. The mandible (fig. 141) resembles generally that of the Argentine Lepidocampa figured by Silvestri (/.c. pl. 6, fig. 4), having an acuminate condyle and four prominent apical teeth, three of which bear minute subsidiary teeth ; on the inner face of the third tooth are a series of ridges, forming apparently a grinding area. The lacinia to which Silvestri drew attention is conspicuous, consisting of a delicate “comb” of eleven teeth springing from a wide base attached to a blunt outgrowth of the inner edge of the mandible just beneath the teeth (fig. 1427). The maaillulae, maxillae, and tongue agree rather closely with the corre- sponding structures in Silvestri’s American Lepidocampa (’99, pl. 6, fig. 6); in our species, however, the lacinial ‘“‘comb” (fig. 143 7) has six processes, each ending in a delicate and slightly inflected lamella. The innermost of these processes has a perfectly smooth inner edge in JL. fimbriatipes, whereas in Silvestri’s figure it is shown with a marginal row of small, sharp teeth. The palp (fig. 143 p) is acuminate, with two or three papillae and stiff spines at its tip; it projects obliquely and inwardly towards the mouth from the galea (fig. 143 9), which is crowned with a group of long bristles, and bears near its anterior terminal edge a blunt, peg-lke sensory structure (fig. 144). The maaxillula (figs. 148, 157 Ml) consists of a sub-triangular lobe with its blunt apex projecting over the tongue (//y) and beset with very fine ridges and hair-like outgrowths. The outer edge of the maxillula is connected with the palp and galea just described, and the arrangement of these parts in Lepidocampa—lying as they do distinctly anterior to the stipes and lacinia— suggests that they belong really not to the maxilla, but to the maxillula, an opinion advocated—after study of the very similar corresponding organs in Campodea and Iapyx—by von Stummer-Traunfels (91) and Hansen (’93). But the base of the galea is clearly connected with the maxillary stipes, the lacinia in insect maxillae generally lies behind the galea, and Borner (08), after careful comparison of these structures in lapyx with those of Machilis on the one hand, and of the Collembola on the other, is convinced that they are rightly referred to the maxilla. In support of this view, it is noteworthy that in the Machilids—probably as regards their Jaws the most primitive of R,I.A, PROC., VOL. XXXIII, SECT. B, [EZ] 26 Proceedings of the Royal Irish Academy. all insects—the palp of the maxillula is vestigial (see figs. 5, 38, Ml), and it seems unlikely that in the lapygidae and Campodeidae the maxillular palp should be well developed and the maxillary utterly vanished. The labium (figs. 145, 152) is, as Silvestri has pointed out (99, p. 393), closely like that of Campodea, but his figures (/.c. pl. 6, figs. 7, 8) are imperfect in detail. The sub-mentwm (fig. 152 s.m.) lies directly in front of the prothoracic presternum (pst), and the mentwm (m) is reduced in extent. The ovoid protuberances (), covered with sensory spines, are probably rightly regarded by Silvestri as palps, while the small conical processes (/), which Meinert (°65) considered to be palps, are evidently—from their internal position, as Silvestri has seen—laciniae. The broad, bristle-bearing lobes that project behind the mouth (fig. 152 g) must thus be recognized as galeae. The legs in their relative lengths andin the proportions of their segments agree closely with those of Z. Webert. The most remarkable feature of these limbs is found in the beautifully fringed processes—apparently pulvilli (fig. 146 pl)—projecting on either side from the small claw-like empodium (emp) below the strongly curved claws (c/). From Oudemans’ description and figure (790, p. 77, pl. vil, fig. 8) it seems that these pulvilli in our Seychelles insect resemble very nearly those of his species. ‘he pulvillus is a delicate, leaf-like plate, both edges of which bear series of stiff, slightly clubbed bristles projecting as a fringe, those of the outer series being longer than those of the inner. Silvestri’s drawing of the foot of the South American Lepidocampa (’99, pl. 7, fig. 19) represents the whole surface of the pulyillus as covered by a number of rather feeble hairs, an entirely different arrangement from that found in Z jimbriatipes. The first abdominal segment, with its appendages, shows the sexual difference characteristic of Campodea. In the female (fig. 135) the hinder edge of the sternum is furnished with a row of simple bristles, while the stylet (fig. 147) carries long bristles from its base onwards and a few spinose papillae at the tip. It is apparently relatively longer and less blunt than in Silvestri’s species (’99, pl. 7, fig. 12). In the male (tig. 153), the stylet is relatively short and thick, with a cluster of spinose papillae at the tip, while the edge of the sternum bears several rows of spines, those of the two hind- most on prominent glandular papillae. The stylets on the succeeding six abdominal segments (fig. 148) bear each two strong spines at the extremity. Silvestri’s description and figure (’99, p. 393, pl. 7, fig. 14) indicate these stylets merely as “setosi” in the species that he discovered in South America. The exsertile vesicles are conspicuous on the abdominal segments from the second to the seventh, inclusive (fig. 135). When thrust out they exhibit a Carpenrer— The A pterygota of the Seychelles. Q7 stiff cylindrical base, bearing the somewhat granulated, bladder-like extremity (fig. 154 ¢.v.). Neither Oudemans nor Silvestri mentions the reproductive organs of Lepidocampa; it is gratifying, therefore, to find that the Seychelles specimens afford material for at least a preliminary account of them. As might have been expected, they resemble rather closely those of Campodea, as described and figured by Grassi (’88. pls. ivand v, figs. 46, 50) and Meinert (65, pl. xiv, fig. 13). In the male the hinder edge of the eighth abdominal sternum (fig. 154, viii) projects as a sub-triangular process, bearing series of long and short bristles, and concealing the external reproductive organs. These are exceedingly simple, consisting of two flattened chitinous genital plates (fig. 155 y.p.) with their free edges sub-semicircular and bearing series of bristles; between these plates the ejaculatory duct opens, so that the whole structure forms a kind of penis, as it is called by Meinert. The very short median ejaculatory duct (fig. 155 d.e.) is formed by the union of paired vasa deferentia (v./.) ; its outer coat is thrown into a series of corrugations, showing that the organ in the specimen examined is in a retracted condition, the retraction being brought about by the action of muscles running parallel to the general direction of the tube, and originating in the abdominal exoskeleton ; when extended the organ would evidently protrude beyond the hinder edge of the eighth sternum. The female's eighth abdominal segment has the hinder edge of the sternum almost straight centrally (figs. 135, 156). Beyond it project a pair of short, blunt, conical processes, with a few bristles (figs. 135, 156 go) which may reasonably be regarded as the gonapophyses; the vulvar opening is between these and a semicircular genital plate (fic. 156 g.p.) corresponding to the dorsal plate similarly situated in the male. Anterior to these structures, and concealed by the eighth sternum, is the slit-like spermathecal opening (fig. 156 spe). In both sexes the hinder edge of the tenth abdominal tergum (figs. 149, 151) is adorned with a series of simple, bifid, and feathered bristles; beyond it projects the pointed telson (figs. 149, 150 te). he tenth sternum also has its hinder edge beset with varied bristles; it is deeply cleft in the middle line (figs. 150, 154), each half partly concealing an acuminate anal valve (v!) which bears an obliquely arranged series of papillae, whence spring long, flexible bristles. External to these valves may be seen the bases of the cerci (fig. 154 ce), which are unfortunately wanting in all the specimens examined. The contents of the rectum are easily visible in some of the specimens, and afford interesting information as to the food of Lepidocampa. They [H 2] 28 Proceedings of the Royal Trish Academy. consist either of conidia of Helminthosporium' and other fungi, or of frag- ments of insect cuticle and bristles; probably the latter indicate that Lepidocampa behaves as a scavenger rather than as a beast of prey. Order COLLEMBOLA. The describer of Collembola, or “Spring-tails,” from the Seychelles has the advantage of a field almost unworked, as one species only of these insects, Acanthurella Braueri, Borner (06), seems as yet to have been recorded from the Archipelago. That insect is represented in the present collection, and so is a species Avrelsonia thalassophila, Borner (07), described lately from Madagascar. The remaining sixteen species now recorded appear to be all new to science. Among the Collembola there is a marked division into two groups, which Borner (’01) is probably justified in regarding as sub-orders. They are thus characterized :— A. Segmentation of the abdomen well marked, occasionally the fifth and sixth or the fourth, fifth, and sixth segments partially fused. Body elongate in form. ‘I'racheal system wanting (except in the Actaletinae), : : : Arthropleona. B. Abdomen sub-globular with the segmentation obliterated. Tracheal system developed, . : : : Symphypleona. Of these two sub-orders, the latter, which marks a higher degree of specialization than the former, appears to be unrepresented in the fauna of the Seychelles, all the eighteen species here recorded being members of the Arthropleona. The Arthropleona may be most naturally divided into two very distinct families, though Borner in his latest classification (130) regards these as “sections” containing three families each, an unwarranted systematic elaboration. A. Prothorax well developed, with definite tergum, bearing bristles. Cuticle usually granulated, Poduridae. B. Prothorax much reduced, its tergum eater eloped. Cuticle not granulated, Entomobryidae. Only a single Seychelles species belongs tothe former of these two families; the remaining seventeen are all Entomobryidae. ‘For the determination of this genus I am indebted to my colleague, Professor T. Johnson, D.Se. Carpunter—The Apterygota of the Seychelles. 29 Family PODURIDAE. Spring-tails of this family are numerous in the great northern continental tracts, and not rare even in the Arctic regions. In tropical countries they appear to be relatively scarcer, and this scarcity is especially noticeable in insular faunas. The single Seychelles species belongs to a sub-family Neanurinae,' characterized by the slender, elongate mandibles and maxillae (see Plate XLV, fig. 1 mnd, mz, fig. 5), the mandibles being without a grinding molar area, and the jaws being often adapted for piercing rather than for biting. NEANURINAE. Neanura sexoculata sp. nov. (Plate XIV, figs. 1-6.) Three ocelli (fig. 1 oc) and a vestigial post-antennal organ (fig. 1 p.a.) on each side of the head. Foot with claw untoothed, empodial appendage (fig. 5 emp) vestigial. Mavxilla (fig. 3) with acute apex and simple delicate process (?palp. fig. 3p). Fifth abdominal segment with intermediate tubercle (fig. 6 ¢6') distinct from dorso-lateral tubercle (fig. 6 ¢d*); each of these 1 Borner (’06, pp. 156-7) proposes to replace the established name (Neanura) of the typical genus of this sub-family by Achorutes, which Templeton gave (Trans. Ent. Soc. Lond., 1836) to a genus comprising two diverse species—(1) dubius (belonging to Achorutes as understood by Tullberg, Lubbock, Schott, and the great majority of modern writers) and (2) muscorum, belonging to Gervais’ Anura, 1841 (modified into Neanura by MacGillivray, 1893). Bérner wishes to revive for the former of these two groups Bourlet’s generic name Hypogastrura (Mem. Soc. Science Agric., Lille, 1839), which is stated by its author to be founded for Podwra aquatica Linné, although the description and figure given show—as Borner correctly points out—that Bourlet had in view a species con~ generic with Templeton’s Achorutes dubius. Hence Borner argues that Hypogastrura must stand as the generic name of this group, and muscorum must become the type of Achorutes, Templeton. Boérner’s argument seems reasonable, and he has been followed in this revision of nomenclature by many subsequent writers. Yet his decision prejudges the question, still under consideration by the International Vommission on Zoological Nomenclature, whether the type of a genus based on a misidentified species ought to be fixed by what the Suthor states or by what he means. In the ‘‘ Smithsonian Inst. Publication,” No. 2256, 1914, pp. 152f., this question is argued by a number of zoologists from opposite standpoints, and is finally reserved by the Commission for con- sideration ; and if this decision hold an author to the letter of his statement, Hypo- gastrura becomes a synonym of Podura. Until, therefore, the principle shall have been settled by authority, I prefer to retain a nomenclature which nobody can misunderstand, for Neanura can mean one genus, and no other. Achorutes, thanks to Borner’s ‘‘ emenda- tion,” has become ambiguous, as any name must when it gets transferred from genus to genus in the same family. Borner himself gives a startling exhibition of the inconvenience and confusion resulting from changes of this kind, by using Achorutes in one sense in the introduction to his paper (1906), and in the other sense in the systematic portion of the same paper ! 30 Proceedings of the Royal Irish Academy. tubercles bearing a sensory bristle dorsally. Abdominal dorso-lateral tubercles rounded. Length 25mm. Colour yellow. Localities.—Mahé: Cascade, 1000 feet (1908, 4 specimens). Silhouette (1908, 2 specimens). Two recently described Oriental species, VV. pudibunda Jmms (12, pp. 86-7), and WV. dubiosa Ritter (1%, p. 397), resemble JV. sezoculata in colour, in the number of ocelli, and in the structure of the foot-claw. Imms and Ritter, however, give no details as to the maxillae and the abdominal tubercles, on which Borner (06, pp. 167-0) has laid stress in distinguishing sub-genera in this genus. But in WV. pudibunda Imms, a specimen of which I have lately had an opportunity of seeing, the intermediate tubercles of the fifth abdominal segment are fused with the dorso-lateral tubercles as in the European JV. muscorum Temp!. The simple maxilla, the distinction of the intermediate from the dorso-lateral tubercles of the abdominal terga, and the presence of sensory bristles on each of these tubercles in J. sevoculata combine to place the species in Bérner’s sub-genus Lobella, founded for the reception of a Japanese inseet—WNeanwra (Lobella) Sauteri. Borner, however, describes the dorso-lateral tubercles of the four anterior abdominal segments as “ zapfenartig” in Lobella; in our Seychelles insect they are rounded like the other abdominal tubercles, so that in this character an approach to typical Neanura (NV. muscorum Templ.) is shown. The main features of the head and its appendages may be seen by reference to the drawings (figs. 1-3). The cuticle is covered with strong granulations, and ears three prominent sub-hemispherical bristle-bearing tubercles on each side. The ocelli (fig. 1 oc) are imperfectly defined: two lie in front closely apposed, and one behind. In front of the two ocelli is a smooth, cuticular area surrounded by strong granulations; this seems to represent a vestigial, post-antennal organ (fig. 1 p.a.). The feelers are of the short, stumpy build usual in Neanura;at the tip of the terminal segment (fig. 2) may be seen retractile sensory papillae, near which are some sensory bristles and short spines. The mandible (fig. 1 mnd) is long and slender ; its proximal end evenly rounded, its tip blunt and toothless. ‘he mawilla (tig. 3) has a simple needle-like apex, near which is attached a delicate, pointed process (p), which may represent the palp. The labium (fig. 4) consists of paired elongate plates with somewhat serrate edge, borne on a median sub-triangular sclerite. The foot and its claw are of the type usual in the genus; no tooth can be seen on the claw, but a minute slender vestige of the empodial appendage (“inferior claw” of older authors) may be distinguished (fig. 5 emp.). The Carprnter— The Apterygota of the Seychelles. 31 arrangement of the tubercles of the abdominal segments and their sensory bristles has already been sufficiently described. The spinose bristles, characteristic of Neanura, are in this species numerous and prominent on all parts of the body. Family ENTOMOBRYIDAE. Except for MNeanwra sexoculata, described above, all the Seychelles Collembola belong to this large family, the relationships of whose numerous genera have formed the subject of much discussion among specialists. General agreement exists as to the recognition of three principal sub- families, one of which—the Tomocerinae—is not represented in the present collections. In his latest classification of the Collembola, Borner (’13)) proposes to raise these groups to the rank of families. The most natural definition of these sub-families, as they may more reasonably be regarded, seems to -be that adopted by Schaffer (97), and by Borner in his earlier works (e.g. '03), and their essential superficial characters may be tabulated thus :— A. Fourth abdominal segment equal, or almost equal, in length to the third. Scales wanting. Feeler with third and fourth segments simple and sub-equal in length. Post-antennal organ usually present. Dentes of spring without spines, . 0 . Lsotominae. B. Fourth abdominal segment shorter than the third. Body scaled. Feeler with third and fourth segments ringed, the former much the longer. Post-antennal organ wanting. Dentes of spring spinose, : : : ; : ‘ 0 ; Tomocerinae. C. Fourth abdominal segment usually much longer than the third. Body scaled or unscaled—always scaled if the third and fourth abdominal segments are sub-equal in length. Post-antennal organ wanting, . : : ‘ ; ; . Hntomobryinae. Borner in his later writings (’06, &c.) transferred from the Isotominae to the Entomobryinae a group including the common European Jsotomurus palustris, because these insects bear on the second, third, and fourth abdominal segments sensory bristles or “ bothriotricha,” which are characteristic of the latter, but not of the former, sub-family. Now Isotomourus (of which there is a Seychelles species) resembles typical Isotoma and its allies so closely in all the main points of structure that Borner was obliged to call in a theory of “convergence” to account for the likeness. Unfortunately almost all writers on Collembola during the last ten years hastened to accept Borner’s new classification, although his “ bothrioticha ” (one is figured on Pl. XLV, fig. 18) are far too slender to carry the weight which he assigned to them. Why 32 Proceedings of the Royal Trish Academy. should their presence or absence be regarded as of such moment, when species with or without a post-antennal organ may be left peacefully side by side in the same sub-family ? And now, in his last paper (136), Borner announces the discovery, on the trochanters of the hind legs of Entomobryinae, of another type of microscopical sense-organ. Since these are wanting in the Tsotomurini, he restores this group, although its members possess “ both- riotricha,” to its natural position among the Isotominae. I am glad, therefore, that before reading Borner’s latest “system” I had decided to be unfashionable, and to retain Isotomurus and its allies among the Isotominae, especially in view of certain admissions previously made by Borner with regard to Axelsonia, an allied genus of very great interest discussed below. TSOTOMINAE. Two species from the Seychelles are referable to this sub-family as just defined, both belonging to the disputed group of the Isotomurini; the typical Isotomini are apparently absent from the fauna of the Archipelago. ‘The species represent two distinct genera which may be readily distinguished. A. Foot-claw with distinct basal filiform processes. Bothriotricha simple, . . é . a : : : : Axelsonia. B. Foot-claw without basal processes. Bothriotricha feathered, Jsotomurus. Axelsonia Borner. This genus was diagnosed by Bérner (’07, p. 147) for a marine species found in barnacle-shells on the Manavara reef off the coast of Madagascar. He had, in the previous year ('06, p. 149), published the name, referring to the genus in addition to the Malagasy species, Zsotoma mitida Folsom (799 a, p. 264, figs. 14-18), from Japan. The slender claw-processes (fig. 12 Lp.) and the simple bothriotricha on the abdominal segments serve to distinguish most definitely Axelsonia from all known Isotomine or Isotomurine genera. Axelsonia thalassophila Borner. (Plate XIV, figs. 7-14.) This species was founded by Bérner (/.c., pp. 147-150, figs. 1-7) for marine spring-tails collected in barnacle-shells on a reef in Antongil Bay (east coast of Madagascar). The Axelsoniae of the Seychelles collection do not appear to differ specifically from Borner’s insects; the only noteworthy divergence is in the comparative lengths ot the third and fourth abdominal segments, the former being distinctly the longer in the insects now recorded Carpenter—The Apterygota of the Seychelles. 30 (see fig. 7), whereas in Borner’s specimens these segments are described and figured (p. 147, fig. 1) as of almost equal leneth. The presence of minute secondary segments or “jointlets”” between the second and third and third and fourth antennal segments (see fig. 9) is noteworthy. ‘his feature is mentioned by Folsom in his description of 4. nitida. Locality—Aldabra, from algae, Bassin Cabris, Picard Island. (J. C. F. Fryer, coll. 9th April, 1909.) On account of the exceptional interest of this species, figures are given of the leading structural features of the Aldabra specimens, that it may be seen how closely they agree with Borner’s types. The details shown in Folsom’s drawings of his Japanese species nitida (in which the third abdominal segment is longer than the fourth) agree also very closely with A. thalassophila, and it is possible that we have but forms of one widespread species, with a tropical and sub-tropical range analogous to that of not a few northern Collembola with littoral habitat. The Japanese localities mentioned by Folsom (Tokyo and Niyagi) are apparently, however, not maritime. The geographical relations of shore-haunting insects are especially interesting; a discussion on these is given below (p. 49). In his remarks on Axelsonia, Borner (/.c. p. 150) expressed his opinion that the simple form of the bothriotricha in the genus gives it a position intermediate between the characteristic Isotomini and the Isotomurini. He adds: “Ob sie ein Glied der Isotomini C.B..oder der Isotomurini C.B. ist, lasst sich schwer entscheiden und bleibt inéglicherweise stets dem subjektiven Ermessen [!] der Forscher tiberlassen.” After this admission it is somewhat surprising that he continued to argue for the decisive evidence of the bothriotricha as a character for placing the Isotomurini along with the Entomobryinae. Further on he pointed out that from the presence of bothriotricha in certain Poduridae and Sminthuridae, “ist ihre Entstehung zur Zeit der hypothetischen Protocollembola hochst wahrscheinlich.”’ Surely if this be so, and the presence or absence of these bristles is useless as a family character, it should not have been allowed to override the many important and conspicuous features of structure which led systematists, until a few years ago, to include Isotomurus and Axelsonia in the compre- hensive genus Isotoma. While objecting to Bdérner’s classification as unwarranted and highly inconvenient, I felt in full agreement with him in regarding Axelsonia as “tief an der Wurzel des ganzen Entomobryenstammes ”’ —an additional argument indeed for considering it to be nearly related to typical Isotoma. And now the discovery that these spring-tails have no “trochanteral organs”’ leads Borner, as stated above, to replace the insects where this obvious relationship is emphasized. R.I,A. PROC., VOL. XXXII, SECT, B, (F] 34 Proceedings of the Royal Trish Academy. Isotomursus Borner. This genus was founded by Borner (03, p.171) for the common European Isotoma palustris (Miller), and no other species seems as yet to have been included in it. Members of the genus resemble the typical Isotomini in all important points of external structure, but are distinguished by the presence of the six pairs of bothriotricha already mentioned on the abdominal segments. If we are justified in regarding the possession of these as an archaic character, it is of especial interest to find that Isotomurus is represented in the fauna of the Seychelles. Isotomurus obscurus sp. nov. (Plate XIV, figs. 15-19.) Feelers half as long again as the head (fig. 15); relative length of segments as 3:8:10:11. Ocelli eight on each side, the inner posterior one markedly smaller than the rest; post-antennal organ broadly ovate, only slightly longer than the diameter of an anterior ocellus (fig. 16). Foot with simple, untoothed claw and lamellate acuminate empodial appendage; no tenent hair (fig. 17). Length of third and fourth abdominal segments equal; fifth and sixth abdominal segments distinct, and bearing a few long bristles. Spring somewhat short and stout; dens half as long again as manubrium, tapering rapidly to tip; muecro (fig. 19) with prominent apical and dorsal teeth, and a minute accessory dorsal tooth. Length 175mm. Colour deep violet-blue, except for the sutures of the body-segments and the dens and mucro, which are white. Loculities.—Silhouette, 1908 (6 specimens). Mahé: Forét Noire, 1000 ft., 1908 (2 specimens). This species is closely allied to the European J. palustris (Muller), but the latter has a tooth on the empodial appendage of the foot, and a small ventral tooth on the muero of the spring. Its post-antennal organ also differs from that of the Seychelles species, being narrowly elliptical. J. palustris has been recorded from Java by Borner (06, p. 173) and from Calcutta by Imms (12, p. 93). ENTOMOBRYINAE. From the number of species included in this sub-family, it may be regarded as the dominant group among the Seychelles Collembola. Borner and other systematists recognize several tribes which may be regarded as natural assemblages of genera, and are at least convenient for purposes of classifica- CareenteEr—The Apterygota of the Seychelles. 35 tion. The tribes and genera comprised in the Seychelles fauna may be distinguished thus :— A. Sixth abdominal tergum elongate and cerciform, feelers with six segments, the two distal ringed, . Heteromuricine. including one genus, -Heteronvuricus. B. Sixth abdominal segment normal, not elongate. Feelers with the four segments all unringed. i. Dentes of spring flexible, with dorsal edye corrugated, mucro relatively short and slender, with dorsal and terminal teeth and a dorsal spine, . . Entomobryini. m. Body unscaled, 6 : : : Entomobrya. n. Body scaled. Mesonotum prominent. o. Dentes of spring without spines, Lepidocyrtus. p. Dentes of spring with spines, . Acanthurella. i. Dentes of spring rigid, without corrugations along dorsal edge. q. Muero of spring relatively short and broad, ; : Paronellins. s. Body scaled, feelers shorter than body, dentes with spines, . 0 . . Microparonella. t. Body unscaled, feelers longer than body, dentes unspined, Cremastocephalus. r. Mucro of spring elongate—usually a third or a quarter as long as dens, which carries series of large ribbed scales, . : : . i ; Cyphoderini. genus, Cyphoderus. HETEROMURICINI. Heteromuricus Imms. This remarkable genus was established by Imms (12, p. 92) for a species A. cercifer, found under dead leaves at Calcutta. The name was given on account of the conspicuous tail-process at the hinder end of the abdomen ; this, however, appears not to be a “median cercus,” as Imms suggested, but the elongate tergum of the sixth abdominal segment. Imms has referred this insect to a new sub-family, the Heteromuricinae, intermediate between the Tomocerinae (seemingly unrepresented in the Seychelles fauna) and the Entomobryinae. He suggests, however, its probable affinity with Heteremurus (Wankel), which is regarded by most recent students as a member of the [F 2] 36 Proceedings of the Royal Irish Academy. Orchesellini. The five- or six-segmented feelers of Heteromuricus show a correspondence with Orchesella, but this is a character that cannot be regarded as of great importance, and the mucro of the spring (fig. 24), with its two teeth and a slender inclined spine, is that of a typical Entomobryine. The jaws, however (see below and next page), are in many respects like those of Orchesella. Until further allied genera shall have been discovered it is, perhaps, best to retain Heteromuricus as the unique representative of a distinct tribe. The insects have a close superficial likeness to species of Tomocerus, on account of the ringed feelers. From these, however, the form of the foot and mucro, and the absence of spines on the dentes, distinguish them at once. Moreover, in Tomocerus the third abdominal segment is longer than the fourth, whereas in Heteromuricus the fourth is clearly longer than the third. In the Seychelles collection a new species of this genus is represented by numerous examples, which differ in many respects from H. cercifer Imms. Heteromuricus longicornis sp. nov. (Plates XV, figs. 20-24; XVIII, figs. 74-77.) Feelers (fig. 20) nearly as long as the body, six-segmented; the proportional lengths of the segments as 1:8:3:9:32:14, the fifth and sixth segments (except for the proximal end of the former) ringed and surrounded with whorls of short, stout bristle (fig. 22). Eight ocelli on each side of the head, the two posteriors of the inner row very small (fig. 21). Legs with scales and feathered hairs; a single filiform bristle near the tip; claws with minute internal teeth near the base, empodial appendage slender, lanceolate, untoothed (fig. 23). Spring half as long as the body; dens 1} times length of manubrium (fig. 20), bearing long feathered hairs; mucro with evenly curved terminal and dorsal teeth, and a fine dorsal spine (fig. 24). Length 3°5 mm. Colour of sealing dark slaty-grey. Localities—Silhouette (1908, 4 specimens). Mahé: Forét Noire, 1000 ft. (September, October, 1908, many specimens). This species is very easily distinguished from H. cercifer Imms, which has the feelers only about half as long as the body, the eight ocelli all about the same size, and the foot-claws with distal teeth. H. Jongicornis also is apparently half as large again as H. cercifer. Opportunity has been taken from the number of examples of this insect in the collection to make a study of the jaws, as the details of structure known about the genus are scanty. There is nothing remarkable about the mandible (Plate XVIII, figs. 74-75), except that the right one has, just proximal to the apical tooth, four small teeth (fig. 74a), while the left CarPENteR— The Apterygota of the Seychelles. 37 (fig. 75a) has two large teeth, with a couple of small rounded tubercles between them. The maxillulae and tongue (fig. 76) are like those of Orche- sella as described and figured by Folsom (990). The mazillula (fig. 76, Mzl) has its distal free end angular, with two sub-acute prominences; along the proximal region of its inner edge is the usual row of denticles—the distal few being blunt, the rest relatively long and sharp. The tongue (fig. 76 hy) is— like that of Orchesella (Folsom, 998, pl. 3, fig.23)—broad distally, with rounded edges and dorsal depressions bounded by toothed ridges; the supporting foot of the tongue (fig. 76 pd) is strikingly like that of Orchesella. So are the cardo and stipes of the maxilla (fig. 76 ¢. st.), the sub-cylindrical galea (fig. 76 g)—according to the highly probable and ingenious interpretation of Borner ('08)—and the vestigial palp (fig. 76 p), with its long, acuminate bristle, being of the usual Collembolan type. The head of the maxillary lacinia (fig. 77) has three strong external teeth—regarded by Folsom and most students as representing the galea—four lamellae (fig. 77 /m), composed of closely approximated bristles, and a conspicuous “ brush ” (67), consisting of an axis, with lateral filaments resembling an ostrich-plume in appearance. This structure certainly corresponds with the brush of the maxillary lacinia in Corethromachilis, described above (pp. 4, 6, Pl. I, figs. 7-8), and in the “Mittelanhang” figured by Borner (08, Pl. VII, fig. 12) as present in the lacinial head of Tetradontophora, which it resembles rather closely. Food material is visible in the intestines of several examples of this species. It consists entirely of fungus hyphae, and conidia, the latter more fragmental than in the rectum of Lepidocampa (see above, p. 28), but apparently also referable to Helminthosporium. ENTOMOBRYINI. Entomobrya Rondani. Degecria Nicolet, Lubbock, etc. This widespread genus of spring-tails, characterized by the absence of scales, the great length of the fourth abdominal segment, and the foot with a conspicuous tenent hair, is represented in the Seychelles collection by a single species, which appears to be undescribed. Entomobrya seychellarum sp. nov. (Plate XV, figs. 25-27.) Feelers nearly twice as long as head, proportion of their segments as 3:6:5:8. Fourth abdominal segment three and a half times as long as third. Foot-claw with three teeth (fig. 26). Mucro of spring with the usual terminal and dorsal curved teeth, the dorsal spine slightly bent (fig. 27). Colour cream-yellow with violet markings, comprising a streak along 38 Proceedings of the Royal Irish Academy. each side of the head, a lateral stripe along all the body-segments, paired L—J-shaped markings on the hinder edge of the metatergum, a strong transverse band on the hinder edge of the third abdominal segment, two pairs of longitudinal streaks running forward from the hinder edge of the fourth abdominal segment, and nearly the whole of the fifth and sixth abdominal segments. Length 1:3 mm. Localities.—Mahé: Forét Noire, 1000 ft. (October, 1908); Cascade, 1000 ft. (1905); Mare aux Cochons, 1000-2000 ft. (January, 1909). Silhouette (1908). It seems hard to find clear structural characters for defining the species of Entomobrya. In the longitudinal violet bands on the fourth abdominal segment, L. seychellarum recalls the European Z. nivalis DG., but the definite transverse band on the third segment serves to differentiate the present insect. Lepidocyrtus Bourlet. This widely spread genus, whose members are scaled, and have the head overhung more or less by the mesothorax, appears to be especially well represented in tropical countries. ‘The Seychelles collection contains no less than seven species—nearly half the total number of Entomobryinae, so that the dominance of the genus is strikingly exemplified. The Seychelles Lepidocryti may be distinguished thus :— A. Mucro slender, with the terminal tooth strongly procurved dorsalwards. i. Feelers not more than twice as long as the head. m. Colour uniformly dark except spring and segmental sutures. Mesonotum rather prominent, JZ. silvestris. n. Colour pale, with dark lateral streaks, thighs and antennal segments with terminal dark rings. Mesonotum very prominent, . . L. annulicorms. o. Colour pale, except for darkening of third and fourth antennal segments. Mesonotum less prominent, . : . : . L, obsewricornis. p. Colour pale, except for lateral patches on fourth abdominal sterna and at tip of hind thighs. Mesonotum rather prominent, . L. stramineus. ii. Feelers much more than twice as long as_ head. Mesonotum very prominent. g. Feelers about three times as long as head, L. Hryerv. 7, Feelers nearly three-quarters as long as body, L. imperialis. Carpenter— he Apterygota of the Seychelles. 39 B. Mucro stout, with the terminal tooth relatively small and not directed dorsalwards, . : ; : ; ? L. Gardineri. All these species have eight ocelli on each side of the head, arranged in the manner characteristic for Lepidocyrtus (Pl. XVI, fig. 44). The empodial appendage of the front foot (fig. 48) is in all cases less elongate than that of the intermediate and hind pairs (fig. 49). Lepidocyrtus silvestris sp. nov. (Plate XV, figs. 28-30.) Mesonotum moderately prominent, twice as long as metanotum; fourth abdominal segment five times as long as third. Feelers rather less than twice as long as the head; proportion of their segments as 5:7:7:4 (fig. 28). Foot-claw with a single minute tooth, empodial appendage narrowly lanceo- late, tenent hair feebly clubbed at the tip (fig. 29). Spring half as long as body ; manubrium stout, equal in length to dens and mucro together ; mucro (fig. 30) narrow, with slender teeth, the dorsal spine long and acute. Length 1l‘lmm. Colour deep violet: only the intersegmental sutures, the spring and the tips of the feet pale. Locality.—Mahé: Forét Noire, 1000 ft. (October, 1908). Of all described Lepidocyrti known to me, this species comes nearest to L. caeruleus Ritter (12, pp. 389-390), from Ceylon, in which the feelers are proportionately shorter, and the foot-claw has a distal tooth on the inner edge. Lepidocyrtus obscuricornis sp. nov. (Plate XV, figs. 31-33.) Mesonotum but slightly prominent, less than half as lone again as metanotum ; fourth abdominal segment five times as long as third. Feelers nearly twice as long as head, proportion of their segments as 5:9:9:11 (fig. 31). Foot-claw with two rather strong internal teeth (fig. 32). Spring nearly half as long as body, manubrium stout, as long as dens and mucro (fig. 51) ; mucro (fig. 33) and slightly curved. with very prominent teeth, the dorsal spine slender Length, 2mm. Colour, pale yellow, except for lateral violet specks on fourth abdominal sterna and a violet suffusion on the feelers from the tip of the second segment to that of the fourth. Locality.—Mahé : Cascade, 1000 ft. (1908, many specimens). In its type of colouration, structure of mucro, and the hairy feelers, this species resembles L. scaber Ritter (12, pp. 390-1), from Ceylon; the latter, however, has the empodial appendage broadened at the tip. 40 Proceedings of the Royal Irish Acudemy. Lepidocyrtus annulicornis sp. nov. (Plate XV, figs. 34-36.) Mesonotum very prominent, two and a half times as long as metanotum ; fourth abdominal segment four and a half times as long as third. Feelers half as long again as the head, proportion of their segments as 3:5:8:9 (fig. 34). TF oot-claw with vestiges of teeth, empodial appendage very slender (fig. 35). Spring, three-sevenths length of body, manubrium equal in length to dens (fig. 54) ; mucro narrow and elongate (fig. 36). Length 175mm. Colour yellow, with violet tips to all the antennal segments and the thighs, and lateral violet streaks on the body segments, those on the fourth abdominal long and furcate (fig. 54). Localities.—Mahé: Cascade, 1000 ft. (1908, 2 specimens); Forét Noire, 1000 ft. (October, 1908, 2 specimens). This species is very readily distinguished from other members of the genus on account of its darkly annulated feelers and the dark-blue body-markings, which give it the aspect of an Entomobrya. Lepidocyrtus stramineus sp. nov. (Plate XV, figs. 37-39.) Mesonotum moderately prominent (fig. 37), two and a-half times length of metanotum. Fourth abdominal segment three and a-half times as long as third. Feelers one and three-auarters times as long as head, proportion of their segments as 5:9:10:11. Foot-claw with minute internal teeth (fig. 38). Spring more than half as long as body, dens slightly longer than manubrium ; mucro relatively short, with teeth very strongly procurved (fig. 39). Length 2mm. Colour pale yellow, except for violet patches at end of fourth abdominal sterna and at tip of hind thighs. Loealities.—Mahé: Forét Noire, 1000 ft. (8 specimens, October, 1908). This species is somewhat near ZL. dahlii, Schiiffer (98, pp. 419-420), from Ralum in the Bismarck Archipelago; the latter species has a less acuminate empodial appendage, and a differently arranged pattern of dark- blue markings on its prevailing yellow ground-hue. Lepidocyrtus Fryeri sp. nov. (Plate XVI, figs. 40-42.) Mesonotum very prominent, three times as long as metanotum (fig. 40). Fourth abdominal segment five times as long as third. Feeler nearly three times as long as head, proportion of its segments as 5: 12:17 : 18 (fig. 40). Carpenter— The Aplerygota of the Seychelles. 41 Foot-claw relatively short, with proximal aud distal internal teeth (fig. 41). Spring half as long as body, dens slightly longer than manubrium ; mucro narrow, with teeth somewhat short; dorsal spine straight (fig. 42). Length 1:75 mm. Colour pale yellow, with tips of second and third, and most of the fourth antennal segments, a few lateral spots on the body- segments, and broad bands on the hind thighs deep violet. Localities—Mahé: Mare aux Cochons, 1000-2000 ft. (1 specimen, Jan, 1909). Lepidocyrtus imperialis sp. nov. (Plate XVI, figs. 43-46.) Mesonotum excessively prominent. three times as long as metanotum. Fourth abdominal segment eight times as long as third. Feeler three-quarters as long as body, proportional length of its segments as 13: 17: 26: 28 (fig. 43), Foot-claw elongate and straight, with small proximal and distal teeth; empodial appendage narrowly lanceolate (fig. 45). Spring two-thirds as long as body, the manubrium rather longer than the dens (fig. 43), mucro strong with the dorsal tooth broad and the terminal somewhat flattened (fig. 46). Length 25 mm. Colour pale yellow, with irregular lateral violet streaks on the thoracic and fourth abdominal segments; the tips of the second and third, and almost the whole of the fourth antennal segments violet, also the hind thighs (fig. 4°). Localities—Mahé: Forét Noire, 1000 ft. (2 specimens, 1908) ; Cascade, 1000 ft. (2 specimens, 1908). Silhouette (4 specimens, 1908). Féiicité (6 specimens with feelers rather shorter than those from other islands, February, 1909). This and the preceding species resemble the West African (Cameroon) spring-tail, Z. maximus, Schott (938, pp. 11-13, pl. ii), in the relatively long feelers, the very prominent mesonotum, and the type of coloration. Z. maximus has the mucro like that of Z. Mryert, and the foot-claw and appendage like those of LZ. imperialis; it is considerably larger than any examples of the genus from the Seychelles. LZ. robustws Imms (112, p. 94) from Travancore South India, and Z. pictus Schaffer (98, pp. 4.6 417) from the Bismarck Archipelago, are also allied to this group. 5 Lepidocyrtus Gardineri sp. uov. (Plate XVI, figs. 47-50.) Mesonotum not very prominent, nearly three times as long as metanotum. Fourth abdominal segment seven times as long as third. Feeler nearly three times as long as head, proportion of its segments as 4:9:10:12 (fig. 47). Foot-claw very long, with conspicuous proximal and distal internal teeth, the RTA. PROC., VOL. XXXUI., SECT. B. [G] 42 Proceedings of the Royal Lrish Academy. empodial appendage very long, narrowly lanceolate (figs. 48-49). Spring three-fifths as long as the body, the manubrium as long as the somewhat robust dens (fig. 47). Mucero (fig. 50) stout, with the terminal tooth small, not procurved dorsalwards and the dorsal tooth close to the terminal. Length 25 mm. Colour pale yellow, with the greater part of the second, and the whole of the third and fourth antennal segments, the tip of the hind thigh, the lateral borders of the meso- and metanotum, lateral patches on the fourth abdominal terguin and sterna, and the tip of the sixth abdominal segment deep violet. Localities —Mahé: Cascade, 1000 ft. (many specimens, 1908); Forét Noire, 1000 ft. (many specimens, October, 1908). This is a very interesting species, showing the long feelers, legs, and spring, and the type of coloration found in the two preceding species and their African and Oriental relations, while the relatively short and blunt inuero recalls that of the common European L. lanuginosus, Tullberg. Acanthurella Borner. Acanthurella was established by Borner (’06, p. 176) as a sub-genus of Lepidocyrtus for the species from the Seychelles here recorded and another species from Java. The presence of strong spines on the dentes of the spring, the modification of the mucro, and a characteristic facies differing from that of typical Lepidocyrtus, may warrant the recognition of Acanthurella as of generic rank. Acanthurella Braueri Borner. Apparently this is the only species of Collembola hitherto recorded from the Seychelles, Borner (06, p. 176) describes its structural features, but gives no figures, nor any precise locality. From the number of specimens from all the stations mentioned in the present collection, it appears to be common in the islands of Mahé and Silhouette. Localities—Mahé: Forét Noire, 1000 ft. (1908) ; Cascade, 1000 ft. (Oct., 1908); Mare aux Cochons, 1000-2000 ft. (January, 1909). Silhouette (1908). Numerous specimens from all localities. There is little to be added to Borner’s description of this spring-tail. The feeler is less than twice as long as the head, the proportional length of the segments being as 5:11:11:12. The mesonotum is twice as long as the metanotum, the fourth abdominal segment five times as long as the third. The two posterior inner ocelli are much smaller than the other six (fig. 52). The mucro of the spring is very characteristic in form, having a small, almost straight, terminal tooth, and a broad and very prominent dorsal tooth ; the spine is long and distinctly curved (fig. 54). Carpenrer—The Apterygota of the Seycheiles. 43 PARONELLINI. Microparonella gen. nov. Body scaled. Feelers relatively short, four-segmented with the segments not markedly disproportionate. Legs spinose; foot-claw normal, empodium narrow, tenent hair feebly clubbed. Ventral tube elongate, with large pro- trusible bilobed sac. Fourth abdominal segment from three to four times as long as third. Spring elongate, with slender, rigid (Paronelline) dentes ; dens with a row of strong spines, but without terminal scale-appendage ; mucro of the broad (Paronelline) type, with four or five blunt teeth. Type, Microparonella caerulea (sp. nov.) Seychelles. This genus is of considerable interest, as it combines the distinctive Paronelline characters of the spring with the general aspect of an Entomo- bryine insect. In the inconspicuous feelers and the relatively short fourth abdominal segment, it seems much more primitive than most members of the Paronellini, and illustrates an annectant type of structure that might, perhaps, be expected in some members of the fauna of such islands as the Seychelles. Microparonella—as its name implies—differs from most Paro- nelline genera in the small size of its species. On the whole, it comes nearest to Dicranocentroides, Imms (12, p. 102), founded on a North Indian species from the Himalayan foot-hills, with which it agrees in the spinose dentes, but from which it may be readily separated by the much shorter feelers and fourth abdominal segment, and by the smaller size of the insects The two species referred to Microparonella may be distinguished thus :— A. Colour blue, three ocelli on each side, foot-claw with no distal tooth, mucro broader, : ; Z ; : . MM. caerulea. B. Colour yellow, eight ocelli on each side, foot-claw with strong distal tooth, mucro narrow, . . : ; 6 M. flava. Microparonella caerulea sp. nov. (Plate XVII, figs. 55-58.) Feeler twice as long as the head, proportional length of its segments as 5:9:9:12 (fig. 55), Ocelli three only on each side (fig. 56). Fourth abdominal segment three and a-half times as long as third. Foot with slender sensory bristle, no tenent hair; claw (fig. 57) with strong proximal teeth, but without distal tooth. Spring four-fifths as long as body, the dens slightly longer than the manubrium; mucro (fig. 58) twice as long as broad, five prominent teeth, a ventral, two terminals, a dorsal, and a lateral. [G 2] 44 Proceedings of the Royal Irish Academy. Length 1 mm. Colour violet-blue; only the head, the spring, and the segmental junctions yellow. Localities—Mahé : Forét Noire, 2000 ft. (four specimens, October, 1908). Microparonella flava sp. nov. (Plate XVIT, figs. 59-62.) Feeler twice as long as head, proportional length of its segments as 5:8:6:12 (fig. 59). Eight ocelli on each side. Fourth abdominal segment three times as long as third. Foot with slender tenent hair; claw with very prominent proximal and distal teeth (fig. 60). Spring four-fifths as long as hody, dens rather shorter than manubrium; mucro (fig. 62) relatively narrow, with five teeth —a terminal, three dorsal, and a lateral. Length 1 mm. Colour yellow, except for the feeler and a few mottlings on the head and abdomen, blue. Localities.—Mahé : Forét Noire, 1000 ft. (one specimen, October, 1908). This species is evidently nearly related to JZ caerulea, but the structural differences ave quite definite, as well as the colour-distinction. The long, eylindrical ventral tube is characteristic in this genus as in Dicranocentroides Inms; it is conspicuous in both the Seychelles species, and the large bilobed sac is protruded in some of the specimens preserved (figs. 55-61). Cremastocephalus Schott (1897). Pterikrypta, Ritter (1912). First described by Schott for a Mexican species (97, p. 175), this genus is now known to be well represented in the Eastern tropics. It is remarkable among the Paronellini for the absence of scales on the body, the very long feelers, the broadly laminate empodial appendages of the feet, and the presence of a scale-like organ (figs. 66-70) on the dens, close to the base of the mucro. All these characters are distinctive of Ritter’s genus Pterikrypta (12, p. 385), he having apparently overlooked Sch6tt’s paper. The two Seychelles species may be distinguished thus :— A. Muero of spring thrice as long as broad, colour yellow with broad violet bands and streaks, . : c ; BC Sconue. B. Mucro of spring less than twice as long as broad ; colour yellow with narrow lateral violet streaks, : < : C. pallidus. Cremastocephalus Scotti sp. nov. (Plate XVII, figs. 63-66.) Feelers rather longer than whole body (including head and spring), proportional length of antennal segments as 4:9:7:8 (fig. 63). Eight CanpenreR—The Apterygota of the Seychelles. 45 ocelli on each side, the two hind inner ones very small (fig. 64). Fourth abdominal segment four times as long as the third. Foot (fig. 65) with the claw strongly curved, with distinct inner basal teeth and a feeble distal tooth; empodial appendage short and broad, the two supporting lamellae of almost equal length ; tenent hair very long and stout. Spring three-quarters as long as body; dens rather longer than manubrium (fig. 63); mucro relatively narrow, thrice as long as broad (fig. 66), with three terminal teeth, the median very blunt. Length 2mm. Colour pale yellow, with conspicuous violet markings, including a ring at each end of the first antennal segment, a broad ring at the tip of the second, the distal three-fourths of the third, and the whole of the fourth ; broad transverse bands on the mesothorax and the second and third abdominal segments; broad paired, lateral bands on the fourth abdominal segment, and the whole of the last two segments (fig. 63). Locafities.—Mahé: Forét Noire, 1000 ft. (six specimens, October, 1908, ; Cascade, 1000 ft. (six specimens, 1908). This species is nearly related to C. indicus, Imms. (12, pp. 104-5 figs. 58, 59), from Caleutta, but it may be readily distinguished from that, not only by the coloration, but by the much more prominent teeth at the end of the mucro. In C. celebensis Schaffer (98, pp. 407-8) the mucro is narrow, as in C. Scotti, but its teeth are blunt and rounded. As might be expected from the excessive length of the feelers in these insects, very few specimens are perfect. Interesting cases of regeneration, with a reduced number of antennal segments, may often be observed—for example, the two-segmented right feeler shown in fig. 63. Cremastocephalus pallidus sp. nov. (Plates XVII, figs. 67-70, and XVIII, figs. 78-81.) Feelers (imperfect in all specimens) longer than the whole body, first segment may be twice as long as head (fig. 67). Hight ocelli on each side, the two hind inner ones much smaller than the others (fig. 68). Fourth abdominal segment six times as long as third. Foot (fig 69) with claw slightly curved, inner basal teeth distinct, distal teeth obsolete; empodial appendage with outer supporting lamella longer than inner. Spring three- quarters as long as body; dens rather longer than manubrium; mucro (fig. 70) hardly twice as long as broad, with three terminal teeth, the median one truncated. Length 2°5 mm. Colour pale yellow, except for violet streaks on the head and along the edges of the body-segments, and violet patches on the thigh-tips and shins (tiv. 67). 46 Proceedings of the Royal Irish Academy. Localities—Mahé: Forét Noire, 1000 feet (many specimens, October, 1908); Cascade, 1000 feet (many specimens, 1908). This species is distinctly larger than the preceding, and easily recognizable by structural characters as well as by its coloration. Its mucro is rather like that of C. montanus Imms (12, pp. 105-6, fig. 60), from the eastern Himalaya, but in that species the scale-like appendage is quadrate, whereas in both the Seychelles insects that structure is rounded. This latter character is shown also in the American species C. trilobatus, Schott (97, pp. 175-8), which has, however—like C. affinis Folsom (’99 a, pp. 265-6) from Japan—the hairs on the dentes feathered, a condition not found in either of the Seychelles insects, nor apparently in Imms’ Indian species. From Ritter’s somewhat rough figures (12, p. 386), his Pterikrypta sulcata from Ceylon must be very closely allied to C. pallidus, the form of the mucro agreeing almost precisely, and the foot-claws apparently differing but slightly, “eine flache Erhebung,” according to Ritter’s description and figure occupying the place of the sharp basal tooth; the coloration also, as described by Ritter, is strikingly like that of @. pallidus, so that a comparison of types might establish specific identity between the two forms. As no study of the jaws of any member of the Paronellini appears ever to have been made, some account of these structures in Cremastocephalus may be given with advantage (Plate XVIII, figs. 78-81). ‘here is a remarkable general uniformity in these organs throughout most groups of Collembola. In Cremastocephalus the mandible (fig. 78) is of the usual form, and calls for no special remark. The mazillula (fig. 79 Mel) has an acute apex at its inner distal corner, and beneath this a small, blunt lobe; the teeth at the base of the inner margin are somewhat short and strong. In the maxilla (fig. 79) the cardo and stipes are of the usual form; the galea (fig. 72g) has a delicate lobe surrounding its apex, and the palp is very small (fig. 79 p), with a strong, straight bristle. The head of the lacinia (figs. 797, 80, 81) is almost cireular in outline, its three outer teeth (figs. 80, 817) hardly projecting beyond the edge of the evenly rounded lamellae, which are supported by series of radially arranged bristles. ‘The tongue (fig. 79 hy) has conspicuous rounded distal lateral lobes and a pair of strong supporting ridges towards the centre; its foot (fig. 79pd) and the supporting arm (fig. 79br.) of the maxillula resemble those of other genera of the Entomobryidae. CYPHODERINI, The spring-tails of this tribe are blind, white insects, living in under- ground or concealed situations, such as caves, or the nests of ants and termites. Several genera have been described, and a useful synopsis of the CarPenteR— The Apterygota of the Seychelles. 47 group has lately been published by Borner (15a). The great majority of the known species belong to the widespread genus Cyphoderus Nicolet, in which must be placed the only representative of the group in the Seychelles collection. Cyphoderus Nicolet. This genus is characterized by the normal build of the head and the foot- claw (except for the excessively long basal tooth, see fig. 72 4), the presence of a very elongate lamella (fig. 72/m) on the empodial appendage, eleven elongate scales (an inner row of five and an outer row of six), on the dens, and usually the great length of the mucro in proportion to the dens. Cyphoderus insularum sp. nov. (Plate XVII, figs. 71-73. Feeler half as long again as head, relative length of segments as 2:5:3:6 (fig. 71). Foot (fig. 72) with stout tenent hair, the claw with slender basal tooth (0), proximal teeth hardly apparent ; empodial appendage (em) long, sharp, and curved, with acute basal lamella (/m). Spring half as long as hody; manubrium slightly longer than dens and mucro together ; dens half as long again as mucro (fig. 73), which has a small upturned apical tooth and two stout dorsal teeth. Inner distal scale of dens (lig. 73 sc) from two-thirds to four-fifths the length of mucro. ‘Length 1 mm. Colour white. Localities—Praslin : Vallée de Mai, “swarming among termites—probably Arrhinotermes canalifrons (SjOst.)—in fallen log” (December, 1908). Mahé: Cascade, 1000 feet (three specimens, 1908). In the form of the mucro and its proportion to the long dental scale this species resembles C. bidenticulatus, Parona (see Borner, 13a, p. 277), inhabiting ant and termite nests in Italy and South Africa; as well as C. termitum Wahlgren (06, pp. 19-20), a termitophile from the Sudan, and (still more closely) C. genneserae, Carpenter (15), from a salt spring near Tiberias. In the Seychelles insect, however, the mucro is of excessive length as compared with the dens. As regards the structure of the foot-claw, C. insularwm is like the European C. albinos Nicolet in the absence of internal teeth. NOTE. A full set of the specimens described in this paper is deposized in the British Museum (Natural History). A large number of duplicates are in the Cambridge University Museum, and some—through Prof. Gardiner’s kindness—are in the National Museum, Dublin, 48 Proceedings of the Royal Trish Academy. DISTRIBUTIONAL NOTES. In concluding this account of the Apterygota of the Seychelles, some remarks as to the indications afforded by the geographical range of the species or their near allies may be appropriate. There are thirteen species of Thysanura and eighteen of Collembola recorded from the Seychelles archipelago proper and the neighbouring islands of the Indian Ocean. In the first place, it is necessary to tabulate the distribution of these species within the area itself. I, SEYCHELLES GROUP. M.=Maheé. S.=Silhouette.- P.=Praslin. B.=Bird Isl. F.= Feélicité. THYSANURA. Acrotelsa collaris—M.B. Atelura nana—M. Lepidospora Braueri—M. Lapyez silcestris— M.S. Lepidocampa fimbriatipes—MS. Corethromachilis Gardineri—M.S.P. C. brevipalpis—MS. C. gibba —M.S. LTepisma intermedia—M.F. Isolepisma bisetosa —M. Ctenolepisma longicaudata —F. CoOLLEMBOLA. Neanura seroculata—M.8. Tsotomurus obseurus—MS. Heteromuricus longicornis—MS. Entomobrya seychellarum—M 8. Lepidocyrtus imperialis—MLS.F. L. Gardineri_M.. Acanthurella Braueri—MS8. Microparonella caerulea—M. M. faca—M. Cremastocephalus Scotti— M. C. pallidus —M. Cyphoderus insularis—M.Y. Lepidocyrtus obscuricornis— M. L. silvestris —M. L. annulicornis—M. L. straminevs—M. L. Fryeri— MSF. Eleven of the Thysanura and seventeen of the Collembola are thus known to inhabit the Seychelles archipelago in the restricted sense. Mahé, the largest of the islands, has ten of the ‘hysanura and all the seventeen Collembola ; Silhouette has five hysanura and seven Collembola ; Félicité has two Thysanura (one—Ctenolepisma longicaudata, probably an introduced species—not found in any other island) and two Collembola; Praslin has one of each order: finally, from Bird Island has been collected nothing except the single Thysanuran Aerotelsa collaris, a house-dwelling species, probably introduced by man, Carpunrer—The Apterygota of the Seychelles. 49 Il. Cortivy. THYSANURA. Acrotelsa collaris. III. AMIRANTE AND FARQUHAR GROUPS. THYSANURA. Isolepisma bisetosa. Kagle Isl. (Amirante), Cerf and Providence Isls. Farquhar. IV, ALDABRA. THYSANURA. Tsolepisma bisetosa. Acrotelsa elongata. Acrotelsa collaris. A, Scotti. COLLEMBOLA. Axelsonia thalassophila. The poverty of the exclusively coral groups (Coetivy, Amirante, Farquhar, and Aldabra) in Apterygota is evident from the above lists (11, III, and 1V), and is highly suggestive when compared with the fairly rich fauna of the granite islands of the Seychelles archipelago proper (1}. The only Collembolan found outside these granite islands is Awelsonia thalassophila, from the coral Aldabra group—a reef-haunting insect already known from the east coast of Madagascar. This spring-tail may be regarded as at least a possible subject for “accidental” dispersal across sea-channels. According to Gardiner’s conclusions, the present Aldabra group could never have formed part of any continental tract, and the transport of small insects by sea-birds is not to be dismissed as impossible. Still the presence of delicate shore-haunting animals on separated islets or analogous stations is strongly indicative of former continuity, or at least approximate continuity; and though Aldabra may be “oceanic,’ the ancient existence of continental islands in the vicinity is certain. The presence of Axelsonia in Japan is noteworthy in this connexion, pointing to a former very extensive range of the genus. The Isolepisma and the Acrotelsae found in the Amirante and Aldabra groups may have been introduced by means of human intercourse or commerce, as some at least of these species frequent the neighbourhood of dwellings, or live indoors. The contrast afforded to the scanty representation of the Apterygota on these coral islands by the comparatively rich fauna of Mahé and Silhouette, from one or both of which come all the Machilidae, Iapygidae, Campodeidae and (except for Axelsonia) all the Collembola of the R.I 4. PROC,, VOL. XXXIII., SECT, B. (4) 50 Proceedings of the Royal Irish Academy. collection, is most remarkable. It appears that the wingless insects of these islands are indeed evidence of former land-connexions, pointing, like the granite rocks, to ancient continental conditions. For if Apterygota be conveyed to any extent by “accidental” means of dispersal, how can the total absence of terrestrial Collembola from the coral islets be explained ? Comparison, from the distributional standpoint, of the Seychelles Apterygota with those of other countries is rendered difficult by the incompleteness of our knowledge of these insects in all tropical regions. Especially with regard to Madagasear—with which comparison would be of sreat interest—are the records disappointingly few; only three species of Collembola from the great island are included in Borner’s recent paper ('07), while in Escherich’s account (10) of the Lepismidae only five Malagasy species are mentioned. In the lists below, in which the Seychelles Apterygota are grouped according to their known geographical affinities, actual Specific identity is indicated by an asterisk. GENERA PECULIAR TO THE SEYCHELLES. Corethromachilis (its sub-family group, Machilinae, being widespread) 5 spp. Microparonella (the most primitive genus of a tribe—Paronellini—with circumtropical range) 2 spp. SPECIES OF WIDELY RANGING GENERA WITHOUT EVIDENT GEOGRAPHICAL AFFINITIES (7). Isolepisma bisetosa. Isotomurus obscurus. * Acrotelsa collaris. Entomobrya seychellarum. Lapyx silvestris. Lepidocyrtus annulicornis. L. stramineus. SPECIES WITH ALLIES IN MADAGASCAR (2). Acrotelsa Scotti. *Axrelsonia thalassophila (also Japan). SPECIES WITH ALLIES IN AFRICA (3). * Acrotelsa nana. Cyphoderus insularum. *Ctenolepisma longicaudata. SPECIES WITH ALLIES IN AFRICA AND INDIA (5). Lepisma tntermedia. Lepidocyrtus Fryert. Lepidospora Braueri. LL. imperialis. L, Gardineri, Carprenrer—The Apterygota of the Seychelles. ay | SPECIES WITH ALLIES IN INDIA AND CEYLON (5). Lepidocyrtus silvestris. < Heteromuricus longicornis. L. obsewricornas. Cremastocephalus pallidus. SG, stramineus. SPECIES WITH ALLIES IN MALAYA OR AUSTRALIA (3). Acrotelsa clongata Neanura sexoculata (Australia). Acanthurella Borneri (Malaya). SPECIES WITH ALLIES IN INDIA, MALAYA. OR JAPAN AND IN TropIcaL AMERICA (2). Lepidocampa fimbriatipes Cremastocephalus Scotti (Malaya and America). (India, Japan, and America). No very definite conclusions can be drawn from the facts of distribution just set forth, mainly because, with the incompleteness of our knowledge of the distribution of tropical Apterygota, it would be unwise to lay stress on negative evidence. With regard to the Seychelles archipelago itself, however, the apparent absence of the Symphypleona, the more highly organized sub-order of the Collembola, -and of all Poduridae except one species, is noteworthy, indicating that the islands became separated from the great continental tracts before the majority of genera belonging to those groups had been able to spread far. A somewhat parallel case is afforded by the Apterygote fauna of the Sandwich Islands, from which also the Symphypleona seem to be absent, and the Arthropleona comprise only a single species of Poduridae (see Carpenter, 04) belonging to the same tribe as the Seychellean Neanura sexoculata, but to a more primitive genus, Protanura. It is note- worthy, also, that the dominant genus of Arthropleona in the Hawaiian archipelago, as in the Seychelles, is Lepidocyrtus, and that the other Collembola occurring in Hawaii are an Isotoma and two species of Entomobrya. When the Apterygota of the Malagasy and Mascarene Islands shall have been well worked, there will be doubtless recognized many more species with affinity to Seychelles insects than the two mentioned above. The feature that comes out from the analysis with some clearness is the establishment of faunistic links between Africa, the Seychelles, and India; the range of some of the most remarkable of these, such as Lepidocampa and Cremastocephalus, stretch as far west as South America, and as far east as Java and Japan. [Hf 2] ED Proceedings of the Royal Irish Academy. Similar geographical relationships are shown from the studies of various groups of insects and other terrestrial Arthropoda from the Seychelles, as recorded in the Reports of the Percy Sladen Trust Expedition (Gardiner and others, 07-14). Taking the Coleoptera (beetles) we find that, in his review of the Seychelles Curculionidae (op. c. V (X V1), pp. 393-497), Champion mentions that out of 134 species, 100 are endemic, and the rest introduced, or showing affinity to African, Malagasy, or Ceylonese weevils. In different groups special athnity with the fauna of one or other of these regions is shown by the Seychelles insects. Thus Grouvelle finds (/.¢., pp 93-116) that of the Nitidulidae and Heteroceridae, seven species have affinity with Madagascar, one with Africa, and three with Ceylon and the Malayan Islands. Scott, after study of the Hydrophilidae and Histeridae (¢. ¢, pp. 193-235, pl. 14), finds that the Seychelles species have predominant aftinity to those in Madagascar and Africa, much the same result being apparent among the Adephaga (op. c. IV (XV), pp. 239-262). In the Lamellicornia, however (/. c., pp. 215-239), there are three distinct Oriental relationships to five African or Malagasy. The Oriental tendency is still more marked among the Pselaphidae, of which Raffray records (op. c. V (XVI), pp. 117-138, pl. 10) only one African, and one Indo-African, as compared with four Asiatic and seven Malayan aftinities Turning to the Diptera, we find that the tropical distribution of most families is tov imperfectly known for satisfactory analysis; but with regard to the Tipulidae, Edwards (op. cit. LV (XV), pp. 195-214, pls. 10, 11) reckons ten African against four Oriental species. Among the Lepidoptera, Fletcher (op. cit. IL (XIII), pp. 265-324, pl. 17), dealing with the larger and more conspicuous moths and butterflies, mentions—in addition to many species with a very wide range—-thirteen African and eleven Malagasy and Mascarene, as against four Indian and three Malayan species. On the other hand, Meyrick (op. cif. III (X1V), pp. 263-307), describing the more primitive Lepidopteran groups of the Tortricina and ‘lineina, distinguishes between an “ancient and highly specialized fauna” and “all the rest which might have been sporadically derived from the Indian region, excepting two or three which more probably originated in Africa.” In many cases like the above. the more primitive orders or groups seem to show Oriental, and the more specialized, African affinities. Thus Burr states (¢. ¢., pp. 123-133), of the 1 In these references the first volume no. refers to the Reports of the Perey Sladen Trust Expedition, the second (in brackets) to that of the Trans. Linn. Soc. Zool., series 2. 2 Kolbe (Mitt. Zool. Mus. Berlin, vol. vy. 1, 1910), reviewing the beetle-fauna of the Seychelles before the Sladen Reports were ayailable, dwelt on the predominance of the Oriental affinities. CarpunteR— The Apterygota of the Seychelles. 53 lowly Dermaptera (Harwigs) that the Seychelles species show distinctly more relationship with those of Ceylon than with the African. Bolivar and Ferriere (op. cit. YV (XV), pp. 293-800) point out that all the Seychelles Phasmidae show Indian, Malayan, or Australian affinities. And with regard to the Odonata (dragon-flies)—nine species of which had been collected by Wright (°69)—Campion (op. cit. LV (XV), pp. 485-446) finds that while the Aldabra and Assumption species are typically African, the Seychelles insects are predominantly Indian. Two Seychelles species are also Malagasy, two African, three both African and Indian, and five Oriental, while of the six endemic Seychelles dragon-flies, the three Zygoptera (demoiselles) have Asiatic, and the three Anisoptera (the more robust Libellulidae and Aeschnidae) have African affinities ; most students of the dragon-fhes would probably regard the Zygoptera as a more primitive tribe than the Anisoptera. Now it is noteworthy that the affinities of the Seychelles Apterygota are with Oriental more than with Ethiopian-species, and as the Apterygota must be regarded as the most primitive of insects, the distribution of the allies of our Seychelles bristle-tails and spring-tails agrees well with the results obtained from the study of other groups. The establishment of such faunistic links, afforded by delicate insects like the Apterygota, incapable of flight, and living for the most part in concealed situations, is in full accord with the belief entertained by many naturalists, in the existence of a Mesozoic and early Cainozoie continental area joining the countries and archipelagoes now widely separated by the waters of the Indian Ocean. This subject has been discussed from the geographical standpoint by Gardiner (06, 07-14), who accepts Neumayer’s suggestion of a continuous land tract in Mesozoic times from South Africa by way of Madagascar and the Seychelles to India and Ceylon, besides a wide continent stretching across the South Atlantic from Africa to America. Hirst (op. cit. V (XVI), p. 31) points out that the distribution of the scorpionid genus Lychas “is very suggestive of the former existence of continuous Jand between the Oriental region and the southern part of the African continent. In Cainozoic times the Seychelles archipelago must have formed part of the large insular or sub-continental tracts which then, as is generally agreed, occupied much of the area of the Indian Ocean. These geographical changes would explain how the elements of the Seychelles fauna are partly Oriental and partly African in their affinities. Most ancient of all the inhabitants are the purely endemic animals, or those whose range 1s very wide and discontinuous. Gardiner mentions the serpentine amphibians— the Caecilia—as vertebrate examples of this ancient element. ‘They are matched by such Apterygote genera as Lepidocampa and Cremastocephalus, which tell—unless “accidental” means of dispersal can be supposed to 54 Proceedings of the Royal Irish Academy. account for their presence—of vanished land in the areas now covered by the Atlantic and Indian Oceans. As to the details of these ancient land tracts, much difference of opinion has naturally been expressed. Wallace—as is well known to all students of animal distribution—argued (92) strongly against the theory of a continuous continent across the Indian Ocean, even in Mesozoic times; and among the most recent writers on the subject, Sarasin (710, p. 57) denies the probability of such a land area “in der spiiteren Kreidezeit und im Tertiar.’ Blanford, however, in his deservedly famous address (90, pp. 88-99), while admitting Wallace’s contention that the facts derived from mammalian and avian distribution afford but weak support to the theory, spoke convincingly in its favour from the range of lower vertebrates and of mollusca, as well as from the extent of ancient ocean-basins, as shown by the range of marine fossils. The great majority of modern students of distribution accept without hesitation the principle of such a continent. As examples of the support afforded to the theory by advance along different lines of inquiry may be mentioned Germain’s study (709) of the mollusca of equatorial Africa, and Ortmann’s admirable essay (02) on the distribution of freshwater Decapods. With respect to the fauna of the Seychelles, the question whether the area of the archipelago maintained its latest connexion with Madagascar and Africa, or with India and Ceylon, is of much interest and difticulty. Ortmann (/. ¢., p. 329) maintains that the connexion of Madagascar with India was interrupted before that with Africa, and a similar view is expressed in one of the maps illustrating Gardiner’s paper ('06, p. 325), which shows an early tertiary Afro-Malagasy peninsula, in which the Seychelles are seen near the apex. Germain, on the other hand (09, p. 172), imagines “ une longue peninsule Indo-Malgache qui s’effondra, ne laissant plus subsister au début du tertiare quune chaine d'iles assez rapprochées. Madagascar est complétement isolé et n’auva plus, par la suite, que des communications temporaires avec l'Afrique.” The series of maps given by Gadow (713) seem to support in the main this latter view. In Perceval Wright’s paper (’71) on the flora of the Seychelles, an outstanding feature is the description of a species of Nepenthes. This genus of “ pitcher-plants” ranges from tropical Australia to Madagascar, aud is—as Wright pointed out—unknown in Africa. Here again the Seychelles show attinity with the Oriental rather than with the Ethiopian region. The Apterygota cannot be expected to throw much light on geographical details such as these, for our ignorance of extinct members of the group is very great. The predominance of Oriental species in the Seychelles fauna has already been emphasized, and it is remarkable that, with the exception Carprenter— The Apterygota of the Seychelles. 55 of Oyphoderus insularwm, the distinctively Ethiopian and Malagasy insects in our collection do not come from the Seychelles proper at all, but from the island groups to the west—Aldabra and Farquhar. Our knowledge of the Apterygota of Madagascar and the African continent is, however, far too incomplete to allow us to attach great importance to negative evidence on this aspect of the question. But so far as our present information goes, a study of the Seychelles Apterygota indicates that the latest continental connexion of the archipelago was with India and Ceylon rather than with Africa. That the ancestors of these wingless insects did reach their present habita- tions by means of continuous land-tracts is far more likely than that they were carried over wide seas by winds or on floating objects; and even Wallace, upholding though he does the theory of the permanence of oceanic basins, admits that the Seychelles are not typical “oceanic islands,” and suggests their possible former connexion with Madagascar, Their spring- tails and bristle-tails seem to tell us plainly that they were once joined with India and Ceylon. At the same time, the absence from the fauna of many important groups, the specific distinctness of most of the insects, the presence of such an apparently primitive and annectant genus as Microparo- nella, and the elaboration of such highly modified and beautiful types of structure as are exhibited in the species of Corethromachilis, all support the conclusion that the islands have long been separated from any continental area. REFERENCES TO LITERATURE. ‘90. W. T. Branrorp.—Anuniversary Address of the President [on the Permanence of Ocean Basins]. Proc. Geol. Soc., London, 1889- 90, pp. 3-110 (Gn Quart. Jour. Geol. Soe., xlvi, 1890). ‘01. C. Borner.—Vorlaufige Mitteilungen tiber einige neue Aphorurinen und zur Systematik der Collembola. Zool. Anz., xxiii, 1901, pp. 1-15. ‘03. —— Neue altweltliche Collembolen nebst Bemerkungen zur Systematik der Isotominen und Entomobryinen. Sitzb. der Gesellsch. natur- forsch. Freunde, Berlin, 1903, pp. 129-182. ‘06. —— Das System der Collembolen, nebst Beschreibung neuer Collem- bolen des Hamburger naturhistorischen Museums. Mitt. aus dem naturhist. Museum, Hamburg, xxiii, 1906, pp. 147-188. Collembolen aus Ostafrika, Madagascar und Siidamerica. In Voeltzkow’s Reise in Ostafrika. Wissensch, Ergebnisse, ii. Band, pp. 147, 148, pls. 7,8. Stuttgart, 1907, 07. 56 Proceedings of the Royal Irish Academy. 08. C. Borner.—Apterygota (1) Collembolen aus Siidafrika, nebst einer Studie tber die erste Maxille der Collembolen. In Schultze’s Forschungsreise in westlichen und zentralen Siidafrika, 1903-5. Jenaische Denkschr. xii, 1908, pp. 55-68, pls. vi, Vil, °09. —— Die Beissmandibel der Insekten und ihre phylogenetische Bedeu- tung. Zool. Anz. xxxiv, 1909, pp. 100-125. ‘13a. Neue Cyphoderinen. Zool. Anz. xl, 1913, pp. 274-284. 13d. —— Die Familien der Collembolen. Z0., pp. 315-322. 04. G. H. CarpentEerR.—Collembola in Fauna Hawaiensis ii. Jondon, 1904, pp. 299-305, pl. ix. 13a. —— A new Spring-tai! from Galilee. Jour. and Proc. Asiatic Soe., Bengal (n. s.), ix, 1918, pp. 215-217, pl. viii. 130. —— The Irish Species of Petrobius. Irish Nat., xxii, 1913, pp. 228- 04, K. EscrericH.i—Das System der Lepismatiden (Zoologica xliii). Stuttgart. 10. —— Lepismatiden von Madagascar, den Comoren, und Ostafrika In Voeltzkow’s Reise in Ostafrika. Wissensch, Ergebnisse, ii Band, pp. 535-8, Stuttgart, 1910. 99a. J. W. Fotsom.—Japanese Collembola. Proc. Amer. Acad., xxxiv. 1899, pp. 261-274, pls. 1-3. 99d. —— The Anatomy and Physiology of the Mouth-parts of the Collem- bolan Orchesella cincta. Bull. Mus. Comp. Zool., Harvard, xxxy, 1899, no. 2, 40 pp., pls. 1-4. 13. H. Gapow.—The Wanderings of Animals. Cambridge, 1913. 06. J. Srantey GArpINER.—The Indian Ocean. Geograph. Journ, 1906, pp. 313-352; 454-471. and others.— Reports of the Percy Sladen Trust Expedition to "07-14. the Indian Ocean, vols. i-vi. Trans. Linn. Soe. (Zool.) (2), xii—xvii. 709. L. Germaiy.—Recherches sur la Malacologie de / Afrique équatoriale. Arch. Zool. Exp. et Gén., li, 1909, pp. 1-195. ’88. B. Grasst—Anatomia comparata dei Tisanuri e Considerazioni generali sull’ organizzazione degli Insetti. [I Progenitori dei Miriapodi e degli Insetti. Mem. vii.] Mem. R. Accad. d, Lincei, (4) iv, 1887, pp. 543-606, pls. i-v (French Transl. in Arch. Ital. Biol., xi, 1889). 64. A. H. Hatmay.—lIapyx, a new genus of Insects belonging to the Stirps Thysanura in the order Neuroptera. Trans. Linn. Soc., xxiv, 1864, pp. 441-7, pl. xliv. 93. H. J. Hansen.—Zur Morphologie der Gliedmassen und Mundteile bei Crustaceen und Insekten. Zool. Anz. xvi, 1895, pp. 193-8, 201-12 (Transl. in Ann. Mag. Nat. Hist. (6), xii, 1893, pp. 417-482), 90. CarpenteEr—The Apterygota of the Seychelles. D7 A. D. Imats.—On some Collembola from India, Burma, and Ceylon, with a Catalogue of the Oriental Species of the Order. Proc. Zool. Soc., Lond., 1912, pp. 80-125, pls. vi-xi1. J. Luspock.—Monograph of the Collembola and Thysanura. Ray Society, 1875. F. Mervert.—Campodeae: en Familie af ‘lhysanurernes Orden. Naturhist. Tidsskr. (3), 11, 1864-5, pp. 400-440, pl. xiv. (Transl. in Ann. Mag. Nat. Hist (3), xx, 1867, pp. 361-378.) A. E. Orrmann.—The Geographical Distribution of Freshwater Deca- pods, and its Bearing upon Ancient Geography. Proc. Amer. Phil. Soc., xli, 1902, pp. 267-400. J. T. OUDEMANS.—Apterygota des Indischen Archipels. In Weber’s Zoologische Ergebnisse einer Reise in Niederlandisch Ostindien, Leiden, i, 1890-91, pp. 73-92, pls. vi, vil. W. Rirrer.—Neue Thysanuren und Collembolen aus Ceylon und Bombay. Ann. k. k. naturhist. Hofmuseums, Wien, xxiv (1910-11), 1912, pp. 379-398. F, Sarastn.—Ueber die Geschichte der Tierwelt von Ceylon. Zool. Jahrb. Suppl., xu, 1, 1910, 160 pp. C. ScHAFFER— Hamburger Magalhaensische Sammelreise. Apterygoten. Hamburg, 1897, 48 pp., pls. 1-3. —— Die Collembola des Bismarck-Archipels. Arch f. Naturgeschichte, 1898, pp. 395-425, pls. xi—xil. H. Scuorr.—Beitrage zur Kenntniss der Insektenfauna yon Kamerun. 1. Collembola. Bih. t. K. Svensk. Vet. Akad. Handl., xix, pt. iv, no. 2, 1893, 28 pp., pls. i-vil. North American Apterygogenea. Proc. Calif. Acad. Sci. (2), vi, (1896) 1897, pp. 169-196, pls. xvi-xviu. H. Scorr.—Eight Months’ Entomological Collecting in the Seychelles Islands, 1908-9. Trans. Linn. Soc. Zool. (2), xiv, 1910, pp. 21-39. F. SILvestrI.—Breve Discrizione comparativa di Lepidocampa Oudms. con Campodea Westw. An. Mus. Nacion. Buenos Aires, vi, 1899, pp. 391-6, pls. 6, 7. -—— Materiali per lo Studio dei Tisanuri. Boll. Soc. Entom. Ital., xxxui, 1901, pp. 204-249. Nuovi Generi e Specie di Machilidae. Redia, ii, 1904-5, pp. 4-9. —— Thysanura. Zoolog. Jahrb. Suppl., vi: Plate’s Fauna Chilensis, iii, 1905, pp. 773-806, pl. 38-44. Note sui Machilidae. Redia, i, 1905-6, pp. 325-340. R,I.A. PROC., VOL. XXXIII., SECT. B, [Z] 58 Proceedings of the Royal Irish Academy. °08a. F. Stnvestri.—Materiali per lo Studio dei Tisanuri, viii-xi. Boll. Lab. Zoolog. Portici, ii, 1908, pp. 259-396. 08d. Thysanura in Sjostedt’s Kilimanjaro Meru Expedition. Upsala, xviii (2), 1908, pp. 1!-14, pls. ii, ii. 10. Materiali per lo Studio dei Tisanuri, xii-xv. Bull. Lab. Zool. Portici, v; 1910-11, pp. 72-99. ‘11. —— Contributo alla Conoscenza dei Machilidae dell’ America Settentrionale. Bull. Lab. Zool. Portici, v, 1911, pp. 324-350. ‘13. —— Tisanuri raccolti dal Dr. I. Tragardh nel Natal e nel Zululand, Arkiv. f. Zoologi, vii, 1913, pp. 1-15. ‘91. KR. von SrumMER-TRAUNFELS.—Vergleichende Untersuchungen tiber die Mundwerkzeuge der Thysanuren und Collembolen. Sitzb. k. Akad. Wissensch. Wien (Math.-naturwissensch. Classe) c., 1891, pp. 1-20, pls. i, ii. 03. KK. W. VERHorrr.—Ueber die Endsegmente des Kérpers der Chilopoden, Dermapteren und Iapygiden, und zur Systematik von Iapyx. Nova Acta Leop.-Carol. Acad., lxxxi, 1903, pp. 257-802, pls. xvili-xix. Zur vergleichenden Morphologie und Systematik der Iapygiden. Arch. f. Naturgeschichte, 1904, i, pp. 63-114, pls. iv—vi. ‘10. —— Ueber Felsenspringer, Machiloidea, 4. Aufsatz: Systematik und Orthomorphose. Zool. Anz. xxxvi, 1910. pp. 425-4388. 06. E. WAuLGREN.—Apterygoten aus Aegypten und dem Sudan. Results of Swedish Zoological Expedition to Egypt, &e. No. 15. Uppsala, 1906. 92. A. R. Watiace.—Island Life (2nd edition). London, 1892. 69. E. P. Wricur.—Notes on the Dragon-flies of the Seychelles, with a List of the Species and Descriptions of a new Genus and some new Species, by the Baron E. DE Setys-Lonecuamps. Ann. Mag. Nat. Hist. (4), iii, 1869, pp. 270-277. ——— Contributions towards a Knowledge of the Flora of the Seychelles. Trans, R.J.Acad., xxiv, 1871, pp. 571-578, pls. xxvii-xxx, KS) [eS w CarpenrER—The Apterygota of the Seychelles. 59 EXPLANATION OF PLATES. THYSANURA, I-XIII. PLATE I. Corethromachilis Gardiner. Scaling not shown. . Female, side view. x 5. . Male. Head, front view, showing eyes (e); median ocellus (m. 0.), paired ocelli (p.0.); labrum (/br); and base of right feeler. x 21. #. Region from middle of flagellum of feeler. x 21. 4. Mandible (left) of female. c, condyle; a, apex; m, molar area; Or “I ad, ad', adductor muscles; re, retractor muscle. Viewed from behind. x 50. . Portion of tongue (Ay) with its right peduncle (pe), and right maxillula (Mel); g, galea; J, lacinia; p, palp. From female, front view. x 50. . Left maxilla of male, front aspect: te, tentorium ; ¢, cardo; st, stipes ; J, lacinia; g, galea ; p, palp. x 21. . Right maxilla of female: head of lacinia from behind, showing apical teeth (¢), lamellae (a), lateral teeth (¢), and “ brush” (7) composed of lanceolate spines. x 84. . The same, viewed from within. x 84. . Left maxilla of male: inner edge of galea and apex of lacinia, front view. x 84. (Lettering same as in figs. 7 and 8.) . The same, outer corner of galea. x 84. . A sensory spine from galea. x 460. 2. Part of labium of male with left palp (sm, submentum ; m, mentum ; p, palp; g, lobes of galea; /, of lacinia). x 21. . Apex of labial palp with flattened sensory spines. x 230. . Male: shin and foot of hind+eg. x 21. tS . Coxal process of second leg. x 21. . Coxal process of third leg. x 21. . Tip of foot showing claws and scopula. x 125. . Extremity of a lanceolate spine of the scopula. x 460. Prate II. Figs. 19-26, Corethromachilis Gardinert ; 27-34, C. brevipalpis, Abdominal segments, Scaling not shown. In all figures, sm, sternum; s.¢., sub-coxa; sf, stylet; ¢. v., exsertile vesicle; yo, gonapophysis ; pe, penis. [12] 60 a 0 ISS) SOF lo bw b& Ww bw vw bb to > He» ¢ 44. — I) Or ie ce) Proceedings of the Royal Irish Academy. C. Gardinert. Male. 1st abdominal segment. x 21. do. do. 2nd do. Seale do. do. ord do. x 21. do. do. 4th do. x 21, do. do. 7th do. x 21. do. do. 8th do. x Zi do. do. 9th do. x 21, do. do, Distal parts of hind gonapophysis. x 84. C. brevipalpis. Male. 1st abdominal segment. x 21. do, do. 2nd do. eile do. don sarond do. x 21. do. do. 4th do. x 21, do. do. 7th do. Sail do. do. 8th do. x 21. do, do. 9th do. x 21% do. do. Anterior gonapophysis. x 84. Prave III. Corethromachilis brevipalpis. Scaling not shown. . Female. Head, front view, showing eyes (¢), median ocellus (im. 0.), paired ocelli (p. 0.), labrum (/b;), and base of right feeler. x 21. . Region from middle of flagellum of feeler. x 21. . Mandible (left) of female: ¢, condyle; a, apex; m, molar area. Front view. x 50. . Tongue (hy) and left maxillula (Jfr/), front view ; pe, peduncle of tongue ; Pp, palp; y, galea; /, lacinia of maxillula. From female. x 50. . Left maxilla of female, hinder aspect: c¢, cardo; st, stipes; gy, galea; 1, lacinia; p, palp. x 21. . Tip of galea and head of lacinia of same, showing apical teeth (f), lamella (/7) with lateral teeth (#), and brush (07). x 84. . Part of labium of female with left palp (p); s. m., sub-mentum ; m, mentum; g, galea; /, lobes of lacinia, x 21. 2. Coxal process of second leg of female. x 21. 3, Hind-leg of female: ¢, coxa; py, coxal process; &, trochanter ; fe, thigh; ti, shin; ta, segments of foot. x 21. Tip of foot, showing claws (cl) and scopula (se). x 125. Fig. . Male, side view. x 5. . Head of male, front view, showing eyes (¢), median ocellus (m.o.), Canpenter—The Apterygota of the Seychelles. 61 PLATE IV, Corethromachilis gibba. Scaling not shown. and paired ocelli (p.o.). x 21, . Head of male, side view. x21. (Lettering as in fig. 46.) . Region from middle of flagellum of feeler. x 37. . Left mandible of female: ¢, condyle; a, apex ; m, molar area. x 50. . Tongue (dy) and left maxillula (Jfz/) of female, front view ; pe, peduncle of tongue; p, palp; g, galea; /, lacinia of maxillula. x 50. 51. Left maxilla of female, front view: c, cardo; st, stipes; J, lacinia; g, galea; p, palp. x 21. . Part of labium of female with right palp (p): s. m., sub-mentum ; m, mentum ; g, galea ; /, lobes of lacinia, x 21. . Coxal process of second leg of female. x 21. . Hind-leg of female: c, coxa ; pr, coxal process ; ¢, trochanter; fe, thigh ; ti, shin; ta, segments of foot. x 21. . Tip of foot, showing claws (cl) and scopula (sc). x 125. . Second abdominal segment of male: sm, sternum; s.c., sub-coxa; st, stylet; ¢.v., exsertile vesicles. x 21. . Third abdominal segment of male. x 21. . Sixth do. do. do. x 21. . Eighth abdominal segment of male (right half): go, left gonapophysis. x 21. . Left ninth abdominal sub-coxa (s.c.) and stylet (st), with penis (pe) and right gonapophysis (go). x 21. . Left anterior gonapophysis of male. x 84. . Left hind gonapophysis of male. x 84. PuatE V. Figs. 63-66. Corethromachilis brevipalpis, female. 67-70. Lepisme intermedia. 63. Corethromachilis brevipalpis, female. HKighth abdominal segment, ventral view, showing sub-coxa (s.¢.), stylets (s¢.), and left gonapophysis (go). x 21. 64. Ninth abdominal episterna (epst), sub-coxae (s.c.), showing stylet (st) on left; abductor muscle (abd) and gonapophysis (gv) on right side. x 21. 65. 34. Proceedings of the Royal Irish Academy. Left gonapophysis of eighth abdominal segment, showing (A) attachment and segmentation near base, (B) near middle, and (C) at tip, adductor (ad) and extensor (cat) muscles. x 50. . Two scales from abdominal sub-coxae. x 210. . Lepisma intermedia. Male. Dorsal view. Scaling not shown. x 17. . Right sub-coxa {s.c.) and stylet (s¢) of ninth abdominal segment. x 116. . Tip of thigh (fe), shin (¢) and foot (¢a) of hind-leg. x 116. Sensory bristle from shin of hind-leg. x 580. Pirate VI. Tsolepisma bisetosa. Scaling not shown. Male, dorsal view. x 8. Face (f) and labrum (/br). x 62. 3. A bifid bristle from the face. x 210. Two scales from the body. x 210. . Right mandible, seen from behind : ¢, condyle; @, apex; ve, retractor ; ad. p, posterior adductor, and ad, median adductor muscles. x 62. . Apex (a) and molar area (7m) of left mandible, front view. x 62. 7. Tongue (Ay) and left maxillula (wl) ; front view, ye, peduncle of tongue. x 62. Right maxilla seen from the front: ¢. cardo ; st, stipes ; y, galea; /, lacinia ; p, palp; pr, protractor; ad, adductor muscles; /.m. adductor of lacinia ; g.m. of galea. x 62. 79. Labium with left palp. (p); s.m. sub-mentum; m, mentum; g, yalea; 1, lacinia. x 62. . Terminal abdominal segments of male, ventral view ; viii, 8th sternuin ; s.c. 9th sub-coxa; st, 9th stylet; pe, penis. x 20, . Left sub-coxa of ninth abdominal segment, dorsal view (internal), showing base of stylet with muscles. x 62. 2, Terminal abdominal segments of female, ventral view: vit, viii, zx, stylets of 7th, 8th, and 9th segments ; go. viii and go. iz, anterior and hinder gonapophyses ; s.c. 9th sub-coxa. x 2d. PuatE VII. Acrotelsa elongata, female. Sealing not shown. . Dorsal view. x 8. Left mandible, front view; ad, adductor muscle. x 58. CarPENTER— The Apterygotu of the Seychelles. 63 Fig. 85. Distal region of mandible, showing apex, molar area (7), and bifid bristles. x 168. 86. Right maxilla, from behind: c, cardo; st, stipes; /, lacinia; g, galea ; p, palp. x 58. 87. Extremity of galea and lacinia of maxilla. x 168. 88. Labium showing right palp (p); s. m.sub-mentum; m, mentum; g, galea ; 1. lacinia. x 58. 89. Edge of terminal segment of labial palp, showing sensory papillae (s. p.) x 168. 90. Terminal abdominal segments, ventral view; vii, ix, stylets of 8th and 9th segments: s.c¢. 9th sub-coxa. go, ovipositor. x 21. Piate VIII. Figs. 91-100, Acrotelsw Scotti, female. Fig. 101, A. collaris (Fab.). Scaling not shown. 91. Acrotelsa Scotti. Dorsal view. x 8. 92. Left mandible, front view. x 58. 93. Apex and molar area of mandible. x 210. 94. A feathered bristle. x 210. 95. Terminal abdominal segments, ventral view. vizi, ix, stylets of 8th and 9th segments; s.c. 9th sub-coxa; go, ovipositor. x 21. 96. Part of right maxilla, from behind: st, stipes; /, lacinia; g, galea; p. palp. x 58. 97. Edges of terga of two adjacent abdominal segments, showing scars of dorsal and marginal combs. x 16. 98. Labium, showing right palp (p) : s. m. sub-mentum ; m, mentum ; g. galea ; J, lacinia. x 58. 99. Edge of terminal segment of labial palp, showing two of the sensory papillae (s.p.). x 375. 100. Terminal part of hind leg: je, thigh; ¢, shin ; ¢a, foot-segments. x 58. 101. Aecrotelsa collaris. Terminal segment of labial palp. s.y. sensory papillae. x 58. Puate IX. Lepidospora Brauert Wscherich. Scaling not shown. 102. Female. Ventral view. x 8. 103. Face (7) and labrum (br). x 62, 64 Fig. 104. 105. 106. 107. 108. 109. 110. iil. Proceedings of the Royal Irish Academy. Right mandible, front view: c, condyle; m, molar area; a, apex; 7¢, retractor muscle; ad, adductor muscles. x 62. Distal portion of left mandible, hind view. x 62. Tongue (hy) and right maxillula (mal). x 84. Right maxilla, seen from behind : ¢, cardo; st, stipes; /, lacinia; g, galea ; p, palp; ad, adductor muscles of stipes; pr, protractor of cardo. x 62. Tip and inner edge of lacinia, showing apical teeth and comb-process (c.p.). x 210. Ixtremity of comb-process of lacinia, internal view. x 366. Tip of terminal segment of maxillary palp. s. p. sensory papillae; s.r. annular sense-organ. x 210, Labium: sm. sub-mentum; m, mentum; /. lacinia; g, galea; p, left palp. x62. A. Sensory papillae on terminal segment. x 366. PLATE X. Lepidospora Brauweri Escherich. Scaling not shown. 2. Second leg: cz, haunch; #7, trochanter ; fe, thigh: ¢z,shin ; ta, sesments of foot. x 42. . Tip of terminal foot-segment, showing claws and empodium (em). x 210. . Fifth abdominal sternum with stylets (st) and exsertile vesicles (e.v.). Muscles indicated on left side. x 42. 5. Terminal abdominal segments of young specimen, side view: st 8, st 9, stylets of 8th and 9th segments; s.c. 9, 9th sub-coxa. x 42. }. Tenth tergum of adult male, ventral view, showing base of tail-process (t.p.), and apodeme (ap), and base of cercus (ce) on left side. x 42. . Terminal abdominal segments of female, side view. x 42. . Eighth abdominal segment of female with appendages, left half. Ninth right sub-coxa with stylet and gonapophysis, ventral view. x 42. In figs. 117 and 118: vidi, 8th abdominal sternum; s. c., sub-coxa ; st, stylet ; go, gonapophysis, the numbers 7, 8, 9 indicating the abdominal segments. . Tip of right hinder (9th) gonapophysis, showing series of recurved hooks on inner ventral aspect. x 210. . Tenth abdominal tergum of female. x 42, Carpenrer—The Apterygota of the Seychelles. 65 PLate XI. Lapye silvestris. . Head with left feeler (extended), and prothorax with right front leg, dorsal view; from Mahé specimen. x 58. 122. Head with left feeler (contracted), dorsal view ; from Silhouette specimen. x 58. . Segments from middle region of feeler, showing the passage from contracted to extended condition. x 168. . Fourth, fifth, and sixth segments of feeler, ventral view, showing bothriotricha or “auditory” bristles (s). x 420. . Head, ventral view, showing labium: y. gena; m. mentum; J. lobe (galea and lacinia fused ?); p. labial palp. x 58. . First abdominal sternum of male: sé. stylet; v. vesicle. x 58. . Sixth to tenth (vi-v) abdominal terga of male with open forceps, dorsal view. x 58. . Seventh abdominal sternum of male. x 58. . Eighth abdominal sternum of male with extruded genital plate (9. p.) and gonapophyses (go). x 58. . External reproductive organs of female, ventral view as protruded: g.p. genital plate; 7. lobe; g. gonapophysis; vl. vulva; spe. sper- mathecal opening. x 168. . Papilla at tip of gonapophysis. x 620. . External male reproductive organs, ventral view, as extruded: g. p. genital plate; go. gonapophysis; r. ridge shielding aperture. x 168. . The same, dorsal view; the aperture (y. a.) showing through the translucent genital plate (g.p.). x 168. . Ninth and tenth (tr. x.) abdominal sterna of male with closed forceps. x 58. PuLate XII. Lepidocampa fimbriatipes, female. Scaling not shown except in fig. 148. . Ventral view : go, gonapophyses. x 33. . Right half of head and pronotum, with seven proximal segments of feeler, dorsal view: sw, epicranial suture; s, “auditory” bristles. x 62. 137. One of the “auditory” bristles from feeler. x 590. R.I.A. PROC., VOL. XXXIII., SECT. B. [kK] Proceedings of the Royal Irish Academy. . Antennal organ at apex of terminal segment of feeler. x 590. . Edge of pronotum with pinnate bristles. ~ 210. . A scale from the thorax. x 210. . Right mandible, ventral view. x 92. . Tip of mandible, showing apical teeth: ridged molar area (mo) and lacinia (2). x 580. Left maxilla, right maxillula (J/z/l) and tongue (Hy), viewed from behind (the left maxilla removed to expose the maxillula): ¢, cardo; st, stipes; g, galea; p. palp; pe, peduncle of tongue. x 210. . Tip of maxillary galea, with spines and peg-like sense-organ. x 210. . Labium: inner portion of spinose lacinia (/), and palp (p). x 210. . Tip of hind-foot, showing paired claws (c/.), empodial claw (emp.), and fringed pulvill (p/.). x 370. . Left stylet of first abdominal segment. x 210. . Left stylet of fourth abdominal segment. x 210. . Extremity of abdomen, dorsal view, showing hinder edge of tenth tergum and triangular telson (fe). x 92. . The same, ventral view; the left half of the tenth sternum has been partly removed to expose the anal valve (v/); fe, telson. x 92. Prate XIII. Lepidocampa simbriatipes. Male (figs. 151-5). Female (figs. 156-7). Scaling not shown. . Male, dorsal view. x 33. . Right half of labium, with part of presternum (prst.), and sternum (st.) of prothorax: s. 7m, sub-mentum ; m, mentum; /, lacinia; g, galea ; p, palp. x 168. 3. Sternum with left stylet of first abdominal segment. x 168. . Hinder abdominal segments (vii-x), ventral view: sf, stylet; ¢. v., exsertile vesicles; v/, anal valve; ce, base of cercus. x 62. 55. Male genital ducts and armature: viii, front edge of eighth abdominal sternum; v.d., vas deferens; d.c., ejaculatory duct; g.p., genital plate. x 168. . Female external reproductive organs: viii, front edge of eighth abdominal sternum; y.). genital plate; go. gonapophysis; spc. spermatheca. x 168. . Leit maxillula (M27) with edge of tongue (Hy). x 250. Fig. Ik io i CarPENTER—The Apterygota of the Seychelles. 67 COLLEMBOLA, XIII-XVIII. Phare XIV. Figs. 1-6, Neanwra sexoculata. Left half of head, dorsal view, showing feeler, ocelli (0c), post-antennal area (p.a.), labrum, right mandible (mnd), and left maxilla (mz, dotted in situ), tips of maxillulae (mal), and tongue (hy). ~ 168. . Apex of fourth antennal segment, showing two retractile sense-papillae, sensory bristles, and short spines. x 250. . Right maxilla: p, palp. x 168. . Labium, ventral view. x 168. . Terminal segment of leg, showing claw and vestigial empodial appendage (emp). x 168. . Fourth, fifth, and sixth abdominal segments, dorsal view. On the fifth segment are marked the dorsal (td), the intermediate (¢b'), and the dorso-lateral (¢b°) tubercles. x 62. Figs. 7-14, Avelsonia thalassophila Borner. . Side view. x 46. . Left group of ocelli, with basal segment of feeler. x 210. . End of third and base of fourth antennal segment, with minute inter- mediate “jointlet ”; sensory pegs (s.)».) near tip of third segment. x 210. . A sensory peg from third antennal segment. x 840. . Tip of fourth antennal segment with apical process. x 210. . Extremity of hind foot with claw (c/.), lateral process (/. p.), and empodial appendage (emp). x 370. . Catch, side view. x 370. . End of dens, and mucro of spring. x 370. Figs. 15--19, Zsotomurus obscwrus. . Head, dorsal view, with left feeler. x 62. . Left group of ocelli, with post-antennal organ (p.a.). x 210. . Hind foot, with claw and empodial appendage. x 370. . Abdominal sensory bristle (bothriotrichum). x 370. . End of dens and mucro of spring. x 370. 68 Proceedings of the Royal Irish Academy. PuaTE XY. Figs. 20-24, Heteromuricus longicornis. Fig. 20. Side view. x16. Scaling not shown, but position of prominent bristles indicated. 21. Ocelli of left side, with base of feeler. x 168. 22. A few “jointlets” of third antennal segment, with whorl of feathered bristles. x 210. 23. Hind foot, with claw and empodial appendage. x 168. 24, End of dens, and mucro of spring. x 168. Figs. 25-27, Entomobrya seychellarum. 25. Dorsal view. x 33. 26. Hind foot, with claw and empodial appendage. x 2a0. 27. End of dens, and mucro of spring. x 250 Figs. 28-30, Lepidocyrtus sylvestris. Scaling not shown: 28. Head, feeler, and mesonotum, side view. x 62. 29. Hind foot. 370. 30. End of dens, and mucro of spring. x 370. Figs. 31-33, Lepidocyrtus obscuricornis. Scaling not shown. 31. Side view. x 33. 32. Hind foot. x 250. 33. End of dens, and mucro of spring. x 250. Figs. 34-36, Lepidocyrtus annulicornis. Scaling not shown. 34. Side view. x 33. 35. Hind foot. x 250. 36. End of dens, and mucro of spring. ~ 250. Figs. 37-39, Lepidocyrtus stramineus. Scaling not shown. 37. Head, feeler, and mesonotum, side view. x 33. 38. Hind foot. x 168. 39. End of dens, and mucro of spring. x 16%. CarprenrEer—The Apterygotu of the Seychelles. = Pratt XVI. Sealing not shown. Figs. 40-42, Lepidocyrtus Fryer. Fig. 40, Head, feeler, and mesonotum, side view. x 33, 4]. Hind foot. x 250. 42. Mucro of spring. x 250. Figs. 43-46, Lepidocyrtus impertalis. 43. Side view. x 35. 44. Ocelli of right side of head. x 168. 45, Hind foot. x 250. 46. End of dens, and mucro of spring. x 250. Figs. 47-50, Lepidocyrtus Gardinert. 47. Side view. x 38. 48. Fore foot. x 250. 49. Hind foot. x 250. 50. End of dens, and muero of spring. x 250. Figs. 51-54, Acanthurella Brauerc. 51. Side view. x 33. 52. Right group of ocelli. x 168. 53. Hind foot. x 250. 54. End of dens, and muero of spring. x 259. Pirate XVII. Figs. 55-58, Mieroparonella caerwiea. Scaling not shown. 55. Side view. x 58. 56. Ocelli of left side. x 570. 57. Hind foot. x 370. 58. Terminal part of dens, with mucro. x 370. Figs. 59-62, Microparonella flava. Scaling not shown. 59. Head, with feeler. x 58. 60. Tip of hind foot, with claw and empodial appendage. x 370. 61. End of ventral tube, with protusible sac (#). x 92. 6%. Tip of dens, with mucro. x 370. R.A. PROC., VOL. XXXIII,, SECT. B. [2] 70 Fig. 2 ”o. 68. 69. 80. 81. Proceedings of the Royal Trish Academy. Figs. 63-66, Cremastocephalus Scott. Side view, the right feeler regenerated. x 33. . Ocelli of left side, with base of feeler. x 168. . End of hind foot. x 370. Tip of dens, with mucvo (m) and scale-appendage (a). x 370. Figs. 67-70, Cremastocephalus pallidus. . Side view of head and thorax, with basal antennal segment and fore and intermediate legs. x 33. Ocelli of left side. x 168. End of hind foot. x 370. . Tip of dens, with mucro (m) and scale-appendage (a). x 370. Figs. 71-73, Cyphoderus insularum. Scaling not shown. . Head and feeler. x 33. End of hind foot: }, basal tooth of claw; em, empodial appendage; /m, its lamella. x 370. . Tip of dens from inner aspect and mucro (7m), inner (sc) and outer (sc’) dental seales. x 250. Prats XVIII. Figs. 74-77. Jaws of Heteromuricus longicornis. Right mandible, front view: a, apical teeth; m, molar area; c, condyle. SO: Distal part of left mandible, hind view: a, apical teeth; m, molar area. x 210. . Left maxillula (J£/), tongue (iy), and right maxilla: ¢, cardo ; st, stipes; 1, lacinia; g, galea; p, palp; pd, right foot of tongue, front view. x 210. laf . Head of lacinia: ¢, teeth ; 67, brush; /m, lamellae. x 3 5. Figs. 78-81. Jaws of Cremastocephalus pallidus. . Right mandible, front view: a, apical teeth; m, molar area ; c, condyle. x 370. Left maxillula (J//), with its supporting arm (07); tongue (hy), with its right foot (pd); and right maxilla: c, cardo; st, stipes ; /, lacinia; g, galea ; p,palp. Front view. x 370. Head of lacinia: ¢, teeth; Jim, lamellae. Front view. x 750. Head of lacinia, hind view. = 790. TARO, IR. lo AGNDs, WOlb, ROOMINNG, Seis, 18, PrArnels CARPENTER.—APTERYGOTA OF THE SEYCHELLES, THYSANURA. Il. PLAIE B. XXXII, SECT. VoL. Proc. R. I. ACAv., APTERYGOLA OF THE SEYCHELLES. CARPENTER. THYSANURA. Proc. R. I. Acap., Vor. XXXIII., Sect. B. PLATE III. CARPENTER.—APTERYGOTA OF THE SEYCHELLES. THYSANURA. Proc, R. I. Acap., VoL. XXXIII., Secr. B. PLATE IV. CARPENTER.—APTERYGOTA OF THE SEYCHELLES. THYSANURA,. PLATE V. XXXIII., Sect. B. Proc. R. I. ACAD., VOL. APTERYGOTA OF THE SEYCHELLES. CARPENTER. THYSANURA. Proc. R. I. Acap., VoLt. XXXIII., Sect. B. PLATE VI. CARPENTER.—APTERYGOTA OF THE SEYCHELLES. THYSANURA. PLatre VII. XXXII., Sect. B. Proc. R. I. Acap., Vor. CARPENTER.—APTERYGOTA OF THE SEYCHELLES. THYSANURA. tm ae a * mS) ET alee Poe : z- - Proc. R. I. Acap., VoL. XXXIII., Sect. B. PLATE VIII. z7~V7wam? 92 CARPENTER.—APTERYGOTA OF THE SEYCHELLES. THYSANURA. Proc. R. I. AcAD., VoL. XXXIIL, Sect. B. PLATE IX. 1035 JIT \\AWM \ 106 A CARPENTER.—APTERYGOTA OF THE SEYCHELLES. THYSANURA. Yea Vella Risdon ee nanan etal : ; 5 CR hah at a 4 a _ : jp ae cea sf a Fs . y a5 Ve os ral - Lu ; if - Ss f “ae mi ot | —— = —— Proc. R. I. Acap., Vor. XXXIIL, Sect. B. PLATE X. CARPENTER.—APTERYGOTA OF THE SEYCHELLES. THYSANURA. Proc R. I. Acap., Vor. XXXIII., Secr. B. Prave XT. CARPENLER.—APTERYGOTA OF THE SEYCHELLES. THYSANURA. Proc. R. I. Acap., VoL. XXXIII., Sect. B. PLATE XII CARPENTER.—APTERYGOTA OF THE SEYCHELLES. THYSANURA. Proc. R. EF Acab., VOL. XXXIII., SEcr. B. PLATE XIII Ss =; ry s LP Pe a 7 S : =! fe aes CARPENTER.—APTERYGOTA OF THE SEYCHELLES. THYSANURA. Proc. R. I. Acap., Vor. XXXIII, Secr. B. Prater XIV. CARPENTER.—-APTERYGOTA OF THE SEYCHELLES. COLLEMBOLA. Proc. R. I. Acap., Vot. XXXIII., Secr. B. PLATE XV. ey APEAILAYN AULATENN\\AR TTT CARPENTER.—-APTERYGOTA OF THE SEYCHELLES. COLLEMBOLA. Proc. R. f. Acap., Vor. XXXII, Secr. Tf. Prare XVIII. CARPENTER.—APTERYGOTA OF THE SEYCHELLES. COLLEMBOLA. - Il. A SYNOPSIS OF THE FALSE-SCORPIONS OF BRITAIN AND IRELAND ; SUPPLEMENT. By H. WALLIS KEW. [COMMUNICATED BY PROFESSOR G. H. CARPENTER, M.SC. | Read Marcu 16. Published Juty 7, 1916. CONTENTS. PAGE PAGE Systematic : Distribution : Chelifer (Chernes) Wideri C. L. Koch, Ireland: county records, . . Syl: re-description, . : . 14 | Britain and Ireland: list of species, 82 Chelifer (Chernes) Powellisp. noy., . 74 | Ethology: Chthonius Halberti sp. nov., . ait Habitats, . § : 3 Sear 2G Classified list of species, . : > StncE the publication of “A Synopsis of the False-Scorpions of Britain and Treland” in these Proceedings in 1911 (20), two additions have been made to this part of our fauna: one in Britain and one in Ireland. The former is a Chelifer (Chernes), allied to Chelifer (Chernes) Wideri C. L. Koch, and found in several places in England and Wales; while the latter is a Chthonius, a small and very distinct species, the discovery of which by Mr. J. N. Halbert, M.r.1.a., at Malahide, Co. Dublin, is one of the surprises of recent field-work in Ireland. After a prolonged investigation, both have to be established as new species. Moreover, the known false-scorpion-fauna of Ireland has been enriched, during the last few years, by the addition of several species already known in Britain; and it has been thought useful to include in this communication a brief report on the state of our knowledge of the occurrence of these animals in this country. Finally, a classified list of the British and Irish species is given; and this list has been marked so as to show at a glance the respective faunas of the two countries. 1. A NEW SPECIES OF CHELIFER. In 1913, an opportunity occurred for the examination of abundant material of the false-scorpion here distinguished as Chelifer (Chernes) Powelli. A large number of specimens had then recently been collected by Mr. H. L. R.I.A. PROC., VOL. XXXIII., SECT. B. [M1] 72 Proceedings of the Royal Irish Academy. Powell and myself, at Kirby-le-Soken (Essex), in an old brick-building long used for stabling an ass and containing a quantity of refuse of old hay, ete. ; and a re-examination of my collections showed that it had already occurred in other places: at Ewhurst (Surrey), found by Capt. Aird Whyte, in a cow- shed ; in London, found by Mr. A. J. Chitty, in a granary ; near Lancaster, found by Mr. F. H. Gravely on his own body after camping on a bed of hay; in Eggerslack Wood (Lancashire), found by Mr. R. Standen (8), possibly, as he now thinks, about the nests of small mammals; more recently I found it at Machynlleth (Montgomeryshire), ina stable-loft where fowls nested; and, since it thus appears to be associated with the homes of vertebrate animals, it will no doubt be shown eventually to inhabit hollows of trees and other covered places inhabited by bats, owls, etc. The animal is allied te Chelifer (Chernes) Wideri C. L. Koch, and has been confused with that species, or recorded as C. phaleratus Sim., which is a synonym of C. Wideri.! That species was established by C. L. Koch (1), on Bavarian specimens, found “in dem faulen Holzstaube einer Eiche” ; Simon (3) found it in the forest of Fontainebleau “sous les écorces de chénes”; while in Britain we know it in Sherwood Forest, in the old forest-land of Richmond Park, and in a small remnant of forest at West Wickham (Kent), always under the bark of old oaks; and, unlike other tree-species, it is usually found where the small space between the bark and the wood is choked with a characteristic reddish powdery debris. ‘lhe two species thus appear to have different habitats; and they are certainly distinct. The confusion of the two animals has resulted from the fact that they possess in common a number of good characters which would not be expected to co-exist in more than one species. By these characters they are separated at first sight from all other species of Chelifer (Chernes) represented with us. Together, and no doubt with other species occurring in Europe and as ‘That C. Wideri C. L. Koch, 1843, and C. phaleratus Sim., 1879, are identical is suggested by Simon’s account of his species (3), and is confirmed by paratypes (from Fontainebleau) given to me by him. With regard to Simon’s statement that L. Koch, to whom he had sent specimens, found them to differ from ©. Wideri, the explanation probably is that the comparison was made with the animal found by L. Koch at Nirnberg ‘‘in Hausern”’ (2), probably C. Powelli. It is true that Ellingsen (who has retired from zoological work) regarded C. Wideri and C. phaleratus as separate species ; and, since he stated (7) that the latter was distinguished by less robust palps, it was presumably C. Powelli; and I certainly found that species among specimens labelled “*C. phaleratus” lent to me by Dr. R. Gestro from the Museo Civico at Genoa. But it must not be assumed that all the records of ‘‘C. phaleratus”’ relate to either one or the other of these species : e.g. C. phaleratus(Sim.) Cambridge (10) —_C. scorpioides Herm. ; C. phaleratus (Sim.) Cambridge (5) = C. Panzeri C. L. Koch; C. phaleratus (Sim ) Godfrey (4) = C. dubius (Camb.) ; ete, Kew—A Synopsis of the Fulse-Scorpions of Britain and Ireland. 73 yet imperfectly known, they form a well-marked group within the subgenus. The cephalothorax is of distinctive shape ; as also are the palps, of which the tibia is without marked anterio-proximal convexity ; and an unusually bold granulation is combined with short strongly clavate bristles. Moreover, the bristles of the ventral face of the abdomen (sternites vii.—xi.), instead of being as usual simple or nearly so, are clavate like those of the dorsal face ; and the usual tactile hairs of the extremity of the abdomen (tergite-sternite xi.) are absent. As regards the characters by which the two species are distinguished I refer to the comparative descriptions given below.! It will be seen that the main differences are in the palps; they are sufliciently marked in the palps of the females ; but much more so in those of the males ; a striking fact being that the modifications of the palps of the male take in the two species opposite directions. The palps of the male are more robust than those of the female in C. Widert and more slender than those of the female in C. Powelli. Of other differences, those of the bristle-armature of the abdominal tergites are of practical importance, the bristles of the posterior marginal row being rather closely set in C. Wideri and rather widely set in C. Powelli. The number of bristles in the row is not a fixed character; but it is always greater in C. Widert than in C. Powellr. Taking tergites vi—vill. of the male (the bristles are a little more numerous in the female in both species), the average number in the row is perhaps 19-20 in the first species and 12-13 in the second. The bristles of the sternites are similarly less numerous in the latter species, and thus the difference on the whole abdomen is great. ‘The bristles of C. Wideri are in general somewhat more strongly clavate; and this is particulary noticeable on the sternites. Neither species has a tactile hair on tarsus iv.; but C. Powell: is distinguished from its ally by the presence, about ? removed from the base of the tarsus, of an obtuse bristle longer and less decumbent than the rest. The granulation presents small differences, e.g, that of the palp-femur is noticeably the stronger in C. Widerr. That species, finally, is a little the larger and less deeply coloured; and the “lyriform organs ” of its abdominal tergites (examined in liquid) are less conspicuous. 1 As regards C. Waideri, the description is a revision of that previously given by me (20). That description relates solely to C. Wideri, without admixture of C. Powelli, and the illustration (20, fig. 6) represents the female ; but ‘‘Ewhurst” among the localities belongs to C. Powelli and should be cancelled. In the former paper, and again here, the length of the body is given in millimetres and tenths ; but from the extensible abdomen such measurements, valueless in themselves, merely give an idea of the relative size of the animals. This character of the abdomen has to be borne in mind also when referring to the drawings ; they illustrate primarily the general shape of the palps. Some of the tactile hairs are shown ; but all the bristles are omitted. [M2] 74. Proceedings of the Royal Irish Academy. Chelifer (Chernes) Wideri C. L. Koch. Chelifer Wideri C. L. Koch, 1843 (1). Chelifer phaleratus Simon, 1879 (3). Palps and cephalothorax red-brown, abdominal tergites yellowish-brown ; dull or nearly so, with clavate and toothed bristles. Cephalothorax strongly granulate, both grooves distinct; abdominal tergites granulate with short strongly clavate bristles, those of the posterior marginal row rather closely set, scar-spots moderately distinct, interstitial membrane of dorsum wrinkled almost granulate, bristles of sternites vii.—xi. clavate, tergite-sternite xi. without tactile hairs; galea(?) moderately long, distally with small pro- cesses; palps (?) femur rather robust, as broad as the tibia, three times as long as broad, in front after well-marked elevation from stalk nearly straight, above rising abruptly from stalk, behind the increase is abrupt oblique almost straight with obvious but rounded proximo-posterior corner, beyond this the outline is nearly straight till gently rounded off to the extremity; tibia in front rising gently from stalk and only a little convex, behind beyond stalk very slightly then moderately convex; hand rather narrow nearly parallel descending rapidly to fingers, movable finger slightly longer than hand; anterior margin of fingers with an isolated accessory tooth ; the palp is strongly granulate; bristles of palp short, for most part rather strongly clavate; lower face of maxillae granulate; coxae iv. (?) broad, posterior Fig. 1. Chelifer (Chernes) Wideri C. L. Koch, palp ¢ . margin longer than inner; legs iv. tibia and tarsus without tactile hairs. $ (compared with ? ) with galea poorly developed; palps (fig. 1) a little shorter relatively stouter: femur relatively broader and higher, increasing still more abruptly from stalk, less than three times as long as broad (scarcely 2°8) ; tibia alittle shorter and a little more convex ; hand distinctly shorter relatively broader and slightly higher; fingers closing with wider gape; coxae iv. some- what less broad at base, posterior margin longer in proportion to inner L, 2:4. Under bark of old oak-trees: Sherwood Forest ; Richmond Park; West Wickham (Kent). Chelifer (Chernes) Powelli sp. nov. Palps and cephalothorax red-brown, abdominal tergites horny-brown ; dull or nearly so, with clavate and toothed bristles. Cephalothorax strongly granulate, both grooves distinct; abdominal tergites granulate with short Knw—A Synopsis of the False-Scorpions of Britain and Ireland. 75 clavate bristles, those of the posterior marginal row rather widely set, scar- spots moderately distinct, interstitial membrane of dorsum wrinkled almost granulate, bristles of sternites vii—xi. moderately clavate, tergite-sternite xi. without tactile hairs; galea (? ) moderately long, distally with small pro- cesses ; palps (?) femur less robust than in preceding species, not quite as broad as the tibia, more than three times as long as broad (about 3°3), in front with slight elevation from stalk and beyond middle faintly concave, above rising from stalk less abruptly than in preceding species, behind also the increase is less abrupt obliquely convex without obvious proximo-posterior corner, and beyond this the outline is gently convex to the extremity; tibia in front risimg gently from stalk and only a little convex, behind beyond stalk nearly straight then moderately convex ; hand compared with that of preceding species a little broader at base descending a little more gently to fingers, movable finger about equal to or slightly shorter than hand; anterior Fig. 2.—Chelifer (Chernes) Powelli sp. noy., ¢. margin of fingers with an isolated accessory tooth ; the palp is rather strongly granulate ; bristles of palp short, for most part moderately clavate ; lower face of maxillae granulate; coxae iv. ( ? ) broad, posterior margin longer than inner ; legs iv. tibia and tarsus without tactile hairs. g (compared with ? ) with galea poorly developed ; palps (fig. 2) a little shorter much more slender : femur increasing more gently from stalk, about 3:4 as long as broad; tibia behind beyond stalk and before distal convexity faintly concave; hand narrower and more parallel; fingers not closing with wider gape; coxae iv. somewhat less broad at base, posterior margin longer in proportion to inner. L, 2:3. In old stables, etc., among refuse of hay, etc. ; widely distributed: Surrey, Middlesex, Essex, Lancashire, Montgomeryshire. 76 Proceedings of the Royal Irish Academy. 2. A NEW SPECIES OF CHTHONIUS. The Chthonius here distinguished as Chithonius Halberti was discovered by Mr. J. N. Halbert in 1915 at Malahide, Co. Dublin. Four specimens were obtained : in May, 1 ¢: and in August.2 ¢ 19. They occurred just south of Malahide Inlet, on a rocky patch (limestone) of sea-shore, under stones, the level being below that of the band of orange-lichens and immediately above that of Pelvetia canaliculata, that is to say just above ordinary high-water mark; but the stones would certainly be overflowed by high spring-tides; these stones were resting on moist shelly sand and gravel; and they sheltered, in addition to the Chthonius, Aépus marinus, Gamasid mites, centipedes, insect-larvae, etc. Chthonius is already represented in the Britannic list by Chth. Rayi L. Koch, Chth. orthodactylus (Leach), Chih. tenuis L. Koch, and Chth. tetrache- latus (Preys.): four well-known species, all larger than Chth. Halberti, which is not closely related to any of them. In this genus, as observed by With (6), good specific characters are found everywhere; but especially the teeth of the palp-fingers- provide valuable distinctions; and in this respect Chth. Halberti is entirely unlike the other four. In Chth. Rayi, Chth. ortho- dactylus, and Chth. tenuis the fixed finger has large teeth, well separated from each other, triangular, acute, and inclined backwards: in Chth. tetrachelatus it has similar large teeth, widely separated, triangular, acute, and erect ; but in Chth. Halberti we find the teeth small, in a dense close-set row, parallel and truncate. Among other characters, the cephalothorax is provided with at least 22 full-sized bristles instead of 20 as in Chth. Rayi or 18 as in Chth. orthodactylus, Chth. tenuis, and Chth. tetrachelatus, the posterior margin having at least 6 such bristles instead of 4 as in*Chth. Rayi or 2 as in the other three species. For the rest, the arrangement of the bristles is much the same, except that the median lateral bristle has a position more removed from the lateral margin. Other noteworthy features are the prominence of the median point of the cephalothorax and the small less strongly developed eyes. In the median position of the two tactile hairs of the hand, the presence of bristle-groups on coxae ii. and iii. (and not on coxae i.), and in the general character of the genital opening in the male, this species agrees approximately with the other four. Other species of Chthonius with small palp-teeth are known; but no species has been named from any part of Europe, or indeed from any part of the world, to which the present one can be referred. It is unlikely, how- Kuw—A Synopsis of the False-Scorpions of Britain and Ireland. 77 ever, that the animal is peculiar to Ireland; it may yet be found in Britain ; and it almost certainly occurs in Europe.’ Chthonius Halberti sp. nov. Palps pale horny more or less reddish, chelicerae reddish-horny, cephalo- thorax and abdominal tergites yellowish- horny. Cephalothorax moderately broader in front than behind, with at least 22 full- sized bristles, at least 6 on posterior mar- gin, anterior margin towards middle minutely serrated produced into a well- marked obtuse median point; eyes small, eye i. more than 1 diameter from front, eye li. obscure ; chelicerae robust; palps Chthonius Halberti sp. nov., 3. (fig. 3) femur with 6 bristles in anterio- dorsal row, hand without dorsal depression, fingers much longer than hand (about 1:9), teeth of fixed finger in dense close-set row small parallel and truncate, those of movable finger also in dense close-set row similar but a little lower. L. 1:2. Treland: Malahide, Co. Dublin, 1915; on the sea-shore between the levels of orange-lichens and Pelvetia ; under stones (J. N. Halbert). 3. ON THE STATE OF OUR KNOWLEDGE OF THE OCCURRENCE OF FALSE- SCORPIONS IN JRELAND. These animals have received attention from time to time from several naturalists in Ireland. Specimens have been collected, for instance, on various trips organized by the Fauna and Flora Committee of this Academy, and these specimens have found a home in the National Museum in Dublin. Others have been brought to meetings of the Dublin Microscopical Club, and, in some cases, have been placed in the same Museum. For several years past, moreover, they have been searched for on the excursions of the Belfast Naturalists’ Field Club; and, in particular, Mr. Nevin H. Foster, M.R.I.A., who has had a large share in the organization of those excursions, has collected these animals with enthusiasm in many parts of the country. As 1 Ellingsen (7) has mentioned a Chthonius (Italian specimens) ‘‘ characterized by the very dense row of small teeth on the fixed finger of the palps’’; he referred it to Chth. orthodactylus (Leach), but incorrectly; and it is possible that his animal was identical with ours; on Chth. orthodactylus (Leach) and the specimens of Chthonius (including the type of this species) in Leach’s collection in the British Museum, ef, Kew (20). 78 Proceedings of the Royal Irish Academy. the result of what has already been done it has become possible to draw up a list of fourteen Irish species. This total, which is likely to be somewhat but not greatly increased, includes two species which are unknown in Britain ; but, on the other hand, Britain has ten species which have not yet been found in Treland. Some of these latter, e.g. Chelifer (Chernes) scorpioides Herm. and Chthonius orthodactylus (Leach), will almost certainly be the rewards of future field-work. But others, which are eastern or south-eastern in distribution in Britain, eg. Chelifer Latreillii Leach, Chelifer (Chernes) Wideri C. L. Koch, Chelifer (Chernes) cyrneus (L. Koch), and Chthonius tenuis L. Koch, are probably absent. Of the two main divisions. of the Order— Panctenodactyli and Hemictenodactyli—Iveland is rich in the latter, but remarkably poor in the former. The plan adopted here is that of giving the first known occurrence of each species in the recognized county-divisions, i.e., those of Praeger’s “Irish Topographical Botany.” In the case of all the divisions listed the specimens have been seen byme. When the occurrences have been already made known in print, that fact is indicated by numbers corresponding to those in part ii of the list of books, ete., which appears at the end of this paper. That part of the list is, I believe, a complete bibliography to date of this section of the Trish fauna. 1. Chelifer (Chernes) nodosus Schr.—This species, and the next, occur in manure-heaps, accumulations of garden-refuse, etc., and their range is influenced by man. They have the habit of attaching themselves, by closure of the fingers of one of the palps, to the legs of flies, and they are apt to come to notice in this position in autumn. Down. Downpatrick 1911 (24): on flies’ legs—R. Patterson. 2. Chelifer (Chernes) Godfreyi Kew.—In “The Irish Naturalist,” 1910, p. 138 (19), is a record (headed “Chelifer cancroides,” but relating probably to this species) of two specimens taken in 1908 from legs of a house-fly at Rathmines, Co. Dublin. Specimens taken in 1910, and seen by me, were from the same house at Rathmines. Musca domestica is the fly to the legs of which this species, and the preceding one, usually attach themselves ; but Mr. Halbert recently sent me C. Godfreyi, one of two individuals found in September last on the legs of Stomorys caleitrans, at Glasnevin, Co. Dublin. Dublin. Rathmines 1910 (20, 23): on flies’ legs—N. H. Stephens. Antrim. Belfast 1915: on flies’ legs—J. A. S. Stendall. 3. Chelifer (Chernes dubius (Camb.).—This is a ground species, occurring usually under embedded stones and less commonly under loose stones or among debris. A record by Robert Templeton, in 1836 (9), of Chelifer parasita Herm. ‘‘caught in Island Magee, county Antrim, and at present in Krw—A Synopsis of the False-Scorpions of Britain and Ireland. 79 Mr. Patterson’s cabinet, Belfast,” possibly relates to this species. C. parasita Herm. usually figures as a synonym of C. nodosus Schr., but since the animal from Island Magee had “spatulate hairs” it was not that species; the specimen itself, according to inquiries made in Belfast by Mr. Foster, cannot now be traced. The under-noted record for Co. Wicklow is almost certainly correct; but is marked with an indication of doubt because the single carded and somewhat damaged specimen did not admit of entirely satisfactory examination (16); at the time of its occurrence it was recorded, owing to confusions of nomenclature, as C. phaleratus Sim. (13). Wicklow. Ovoca [Woodenbridge] 1895 (16) ?—Halbert. Antrim. Glynn 1913 \27)—Kew. 4. Chelifer (Chernes) Panzeri C. L. Koch.—C. Panzeri and its diminutive associate Cheiridium museorum are of frequent occurrence in old stables, barns, hay-lofts, ete., and are no doubt dispersed by man. In Britain, but not yet in Treland, they have been found also, apart from man, about old nests of birds, ete. Down. Hillsborough 1915 (27): in a stable-loft —Foster. d. Chelifer cancroides (Linn.).—This species occurs in Britain in old stables, lofts, corn-stores, etc., and has been well established in such places for very many years; but it is doubtfully indigenous. I have not been able to ascertain the conditions in which the I.imerick specimen was found. Limerick. Limerick 1894 (16)—F. Neale. 6. Cheiridium museorum ([each).—This little species has already been referred to under C. Panzeri. It was first recorded for Ireland, but without locality, by Robert 'empleton in 1836 (9). Dublin. Dundrum 1903 (16)—Scharff. Monaghan. Glaslough 1915: in a stable-loft —Foster. Armagh. Poyntzpass 1915: in astable-loft—Foster. Down. Hillsborough 1913 (27): in a stable-loft—Foster. Antrim. ‘Torr 1915: in a hay-loft—Miss M. L. Foster. 7. Obisium (Ideoroncus) Cambridgii (L. Koch.)—0O. Cambridgii is found under stones and among debris, in the open and in woods, and is plentiful where it occurs. In Scotland and England it is western and more or less maritime; but it may possibly be generally distributed in Ireland. Kerry South. Glencar 1911—Kew. Kerry North. Killarney 1911—Kew. Cork West. Inchigeelagh 1907—R. Standen. Antrim. Giant’s Causeway 1913 (27)—C. Oldham. Londonderry. Benevenagh 1912 (26) — Foster. 8. Obisium (Roncus) lubricum (L. Koch.)—This species occurs under stones, R.I.A, PROC., VOL, XXXIII., SECT. B, LN | 80 Proceedings of the Royal Trish Academy. at roots of grass, among moss, ete.; but is rather scarce and difficult to find. In England it is known in the south from Cornwall to Kent, but has not yet been found in the north of England or in Scotland. Mr. Foster’s discovery of the animal in Co. Down—in perfectly natural conditions in Corry’s Glen, Hillsborough—suggests that it may be widely distributed in Ireland. Down. Hillsborough 1912 (25)— Foster. 9. Obisium muscorum Leach.—Found under stones, among dead leaves, moss, etc., in the open and in woods, from sea-level to near the tops of our mountains; widely distributed and abundant. Recorded for Ireland, without locality, by Cambridge in 1892(10) and by Carpenter and Evans in 1896 (11). Kerry South. Kenmare 1899 (16)—Halbert. Kerry North. Jillarney 1893 (16)—Nat. Mus. Col. Cork West. Glandore 1896 (16)—Halbert. Carlow. Fenagh 1909 (18) —Pack-Beresford. Galway West. J.eenane 1897 (16)—Halbert. Wicklow. Ovoca 1894 (16)—Nat. Mus. Col. Dublin. Lucan 1892 (16)—Scharff. Mayo West. Delphi 1897 (16)—Halbert. Donegal East. Bundoran 1913—F ster. Tyrone. Albany 1910—Foster. Armagh. Loughgilly 1893 (16)—W. F. Johnson. Down. Hillsborough 1910 —Foster. Antrim. Kinbane 1897 (16)—R. Welch. 10. Obisium Carpenteri Kew.—This fine species was ascertained in 1909 to inhabit Ireland, having been found in the extreme south-west, at Glengariff, Co. Cork (17). It occurs there on a rocky wooded hill-side, under the flaking outer-bark of Arbutus-trees, in rock-crevices, and among dead leaves (20). The animal is unknown in Britain. Its range with us may perhaps be as restricted as that of Arbutus and Geomalacus; and in this connection, in view of facts known to everyone, it would be of interest to ascertain its continental distribution. As already stated (18), however, confusions of nomenclature leave us little that is certain; but it is perhaps significant that the animal is in France, according to Simon, a southern species, rare in the environs of Paris, and on the contrary commonest of all in Corsica and in Algeria. Cork West. Glengariff 1909 (17)—Kew. 11. Obisium maritimum Leach.—O. maritimum lives on the sea-shore between neap-tide levels, where it is the largest member of an interesting society of terrestrial creatures, inhabiting crevices from which the air is not entirely expelled by the water, It occurs in old deep-seated rock-fissures and Kew—A Synopsis of the Fulse-Scorpions of Britain and Ireland. 81 under large stones lying more or less embedded in permanent resting-places ; and will probably be found all round Ireland. Kerry South. Kenmare Bay 1909 (16)—Kew. Cork West. Bantry Bay 1909 (17)—Kew. Dublin. Malahide 1915—Halbert. Mayo West. Olare Island 1910 (20)—Kew. Down. Ballymacormick Point 1916—Sielfox. 12. Chthonius Halberti sp. nov. This is the animal discovered by Mr. Halbert at Malahide, Co. Dublin, and described in the present paper. It has not yet been found in Britain. Dublin. Malahide 1915—Halbert. 13. Chthonius Rayi L. Koch.—Found under stones, ete. ; widely distributed and abundant.! Kerry South. Kenmare 1909 (17)—Kew. Kerry North. Jallarney 1909 (17)—Kew. Cork West. Glandore 1898 (16)—Halbert. Cork Mid. Corrigrohane 1907 — W. D. Roebuck. Waterford. Waterford 1910—Kew. Kilkenny. Goresbridge 1909— Pack- Beresford. Wexford. Wexford 1910—Kew. Carlow. Borris 1895 (16)—Halbert. Wicklow. Ovoca 1894 (16)—Schartf. Dublin. Lucan 1892 (16)—Scharff. Meath. Beau Pare 1912 (26)—Foster. Longford. Longford 1915—Miss Mabel Murphy. Roscommon. Kiltoom 1910—Stelfox. Sligo. Grange 1913—Foster. Leitrim. Kinlough 1913—Foster. Louth. Clogher Head 1912 (26)—Foster. Monaghan. Monaghan 1912 (26)—Foster. Fermanagh. Castle Caldwell 1912—Foster. Donegal East. Bundoran 1913—Foster. Armagh, Navan Fort 1895—J. N. Milne. Down. Hillsborough 1912—Foster. Antrim. Garron l’ower 1909—Pack-Beresford. 'Tn addition to the references appearing in the text, cf. Carpenter (12, 15) and Foster (22). 2The record (16) ‘‘ Londonderry 1895 (J. N. Milne)’ is a mistake, the specimens having been collected at Navan Fort near Armagh city. LN 2) 82 i Proceedings of the Royal Irish Academy. 14. Chthonius tetrachelatus (Preys.).—Found under stones, ete. ; abundant near the sea and on mountains, and often found elsewhere in old gardens, etc. A record of Chth. orthodactylus (Leach) for the Belfast district (14) relates in reality to this species. Kerry South. Between ‘l'emplenoe and Sneem 1909 (17)—Kew. Kerry North. Galway’s Bridge 1909 (17)—Kew. Cork West. Southern slopes of Esk-mountain near Glengariff 1909 (17)—Kew. Wicklow. Greystones 1911—N. E. Stephens. Dublin. Dublin 1894 (16)—Halbert. Roscommon. Rockingham 1915: ina greenhouse—Miss A. b. Foster. Mayo West. Westport 1909 (21)—Pack-Beresford. Louth. Clogher Head 1912 (26)—Foster. Down. Ballynahinch 1915—Miss M. L. Foster. Antrim. Belfast 1896 (16)—A. G. Wilson. Londonderry. Benevenagh 1913 (27)—Foster. 4, CLASSIFIED LIST OF THE FALSE-SCORPIONS OF BRITAIN AND IRELAND. (* Species found in Great Britain; + species found in Ireland.) PSEUDOSCORPIONES. I. Panctenodactyli. CHELIFERIDAE. Chelifer. 1. Chernes. 1. Chelifer (Chernes) nodosus Schr., 1803. * + 2. Chelifer (Chernes) Godfreyi Kew, 1911. 3. Chelifer (Chernes) Chyzeri (‘l’6m., 1882). = 4, Chelifer (Chernes) scorpioides Herm., 1804. * + 6. Chelifer (Chernes) dubius (Camb., 1892). 6. Chelifer (Chernes) Powelli sp. nov. * 7. Chelifer (Chernes) Wideri C. 1. Koch, 1843. 8. Chelifer (Chernes) Panzeri C. L. Koch, 1836, 2 9, Chelifer (Chernes) cyrneus (L. Koch, 1873). * 10. Chelifer (Chernes) cimicoides (Fabr., 1793). 2. Chelvfer. * 11. Cheliter (Chelifer) Latreillii Leach, 1817. * + 12. Chelifer (Chelifer) cancroides (Linn., 1758). Kuw—A Synopsis of the False-Scorpions of Britain and Ireland. 83 3. Withaus. * 13. Chelifer (Withius) subruber Sim., 1879. Cheiridium. * + 1. Cheiridium museorum (Leach, 1817). Il. He PaiohatomadlanninlD. OBISIIDAE Obisium. 1. Ideoroncus. * + 1. Obisium (Ideoroncus) Cambridgii (L. Koch, 1873). 2, Roneus. 2. Obisium (Roncus) lubricum (L. Koch, 1873). 3. Obisiwm. * =o Se} * + 3. Obisium (Obisium) muscorum Leach, 1817. t+ 4, Obisium (Obisium) Carpenteri Kew, 1910. * + 5, Obisium (Obisium) maritimum Leach, 1817. CHTHONIIDAE. Chthonius. + 1. Chthonius Halberti sp. nov. * + 2. Chthonius Rayi L. Koch, 1873. * 3. Chthonius orthodactylus (Leach, 1817). * 4. Chthonius tenuis L. Koch, 1873. * + 5. Chthonius tetrachelatus (Preys. 1790). 5. List oF Books, PAPERS, ETC., REFERRED TO IN THE TEX. Part I. . Kocu, C. L.—Die Arachniden, x. Nirnberg, 1843. . Kocu, L.—Verzeichniss der bei Niirnberg bis jetzt beobachteten Arach- niden. Abhandlungen der Naturhistorischen Gesellschaft zu Niirnberg, vi. pp. 113-198. Niirnberg, 1877. . Smmon, E.—Les Arachnides de France, vii. Paris, 1879. . GoprrEy, R.—Chernetidea or False-Scorpions of West Lothian. Annals of Scottish Natural History, x. pp. 214-217. Edinburgh, 1901. . CAMBRIDGE, O. P.—On New and Rare British Arachnida. Proc. Dorset Natural History ete. Field Club, xxvi. pp. 40-74. Dorchester, 1905. 84 ~I 10. 19. Proceedings of the Royal Irish Academy. . WirH, C. J.—The Danish Expedition to Siam, 1899-1900. Chelonethi: an Account of the Indian False-Scorpions, together with Studies on the Anatomy and Classification of the Order. Kgl. Danske Viden- skabernes Selskabs Skrifter, (7) iii. pp, 1-214. Copenhagen, 1906. . ELLINGSEN, I!.—Contributions to the knowledge of the Pseudoscorpions from material belonging to the Museo Civico in Genova. Ann. Museo Civico di Storia Naturale, (3) iv. pp. 205-220. Genova, 1909. . Standen, R.—The False-Scorpions of Lancashire and some adjoining counties. Lancashire Naturalist, v. pp. 7-16. Darwen. 1913. Part IT. . TEMPLETON, k.—Catalogue of Irish Crustacea, Myriapoda, and Arachnoida, selected from the Papers of the late John Templeton. Loudon’s Magazine of Natural History, ix. pp. 9-14. London, 1836. CAMBRIDGE, O. P.—On the British Species of False-Scorpions. Proc. Dorset Natural History ete. Field Club, xiii. pp, 199-231. Dorchester, 1892. . Carpenter, G. H., & Evans, W.—A List of Phalangidea (Harvestmen) and Chernetidea (False-Scorpions) collected in the neighbourhood of Edinburgh. Proc. Royal Physical Society, xiii. pp. 114-123. Hdinburgh, 1895. . CARPENTER, G. H.—Zn [Proceedings of] Dublin Microscopical Club. Irish Naturalist, iv. p. 1383. Dublin, 1895. 3. CARPENTER, G. H.—/n [Proceedings of] Dublin Microscopical Club. Irish Naturalist, v. p. 215. Dublin, 1896. . C[ARPENTER], G. H.—/n A Guide to Belfast and the counties of Down and Antrim (British Association Handbook). Belfast, 1902. . C[ARPENTER], G. H.—Zn Handbook to the City of Dublin and the surrounding District (British Association Handbook). Dublin, 1908. . Kew, H. W.—Notes on the Irish False-Scorpions in the National Museum of Ireland. Ivish Naturalist, xviii. pp. 249-250. Dublin, 1909. . Kew, H. W.—A Holiday in South-Western Ireland. Notes on some False-Scorpions aud other animals observed in the counties of Kerry and Cork. Irish Naturalist, xix. pp. 64-73. Dublin, 1910. . Kew, H. W.—On the Irish Species of Obisium ; with special reference to one from Glengariff new to the Britannic fauna. Irish Naturalist, xix. pp. 108-112. Dublin, 1910. SrePHENS, N, E.—Additional record of Chelifer cancroides. Irish Naturalist, xix. p.138. Dublin, 1910. Krw—A Synopsis of the Fulse-Scorpions of Britain und Treland. 85: . Knew, H. W.—A Synopsis of the False-Scorpions of Britain and Ireland. Proc. Royal trish Academy, xxix B. pp. 38-64. Dublin, 1911. . Kew, H. W.—Pseudoscorpiones. Zn Clare Island Survey. Proc. Royal Trish Academy, xxxi. (38) pp. 1-2. Dublin, 1911. . [Fosrer, N. H.J]—Zn Annual Report and Proceedings of the Belfast Naturalists’ Field Club, (2) vi. pp. 367-370. Belfast, 1911. 3. STEPHENS, N. E.—Jn [ Proceedings of] Dublin Naturalists’ Field Club. Trish Naturalist, xxi. p. 25. Dublin, 1912. . CARPENTER, G. H.—Jn [Proceedings of] Dublin Microscopical Club. Irish Naturalist, xxi. p. 117. Dublin, 1912. . Foster, N. H.—Obisium lubricum, a False-Scorpion new to the Irish fauna. Irish Naturalist, xxi. p. 245. Dublin, 1912. }. Fosrer, N. H.—/m Annual Report and Proceedings of the Belfast Naturalists’ Field Club, (2) vi. pp. 588-608. Belfast, 1913. . Kew, H. W.—Jn Annual Report and Proceedings of the Belfast Naturalists’ Field Club, (2) vii. p. 93. Belfast, 1914. . Kew, H. W.—/n Annual Report and Proceedings of the Belfast Naturalists’ Field Club, (2) vii. p. 167. Belfast, 1915. [yl 86 | LI fe DIKETONES DERIVED FROM DIACETORESORCINOL- DIMETHYLETHER. By JOSEPH ALGAR, M.Sc. University College, Dublin. Read May 22. Published Juny 7, 1916. Ix a paper by Ryan and O'Neill (Proc. Royal Irish Acad., xxxii (1915), B, p. 48) two syntheses of diflavone are described. The first consisted of the preparation of dibenzylidene-diacetoresorcinol which was then acetylated and brominated. By this means was obtained the tetrabromide of dihydroxy- dichalkone-diacetate which, when warmed with alcoholic potash, gave diflavone. In the second synthesis diacetoresorcinol-dimethylether was condensed with benzoic ester, and the diketone thus formed was heated with concentrated hydriodic acid. In this manner diflavone was again obtained. Using the first method, attempts were made by Ryan and Algar (Proc. Royal Irish Acad., xxxii (1915), B, p. 185) and by Ryan and Walsh (ibid., p. 193) to obtain other difiavone derivatives from dianisylidene- diacetoresorcinol and diveratrylidene-diacetoresorcinol. The reaction, how- ever, gave dicoumaranone derivatives in each case instead of diflavone derivatives. The present research was undertaken with a view to preparing diflavone derivatives by the second method. Diketones are described which were obtained by the condensation of diacetoresorcinol-dimethylether with anisic, phenyl-acetic, acetic. and oxalic esters. The condensations with anisic and phenyl-acetic esters both gave small yields of the diketone, consequently it was impossible to test the action of hot concentrated hydriodic acid on these diketones. The diketones obtained from acetic and oxalic esters were formed in larger quantities, and the action of concentrated hydriodic acid on these compounds was investigated. In both cases the results were extremely unsatisfactory, the products being either resins or non-crystallisable oils. On prolonged heating of diacetylacetoresorcinol-dimethylether with hydriodic acid and purification of the product, a very small quantity of a colourless substance was isolated, which dissolved in concentrated sulphuric acid, the solution having the strong, green fluorescence characteristic of chromone po. pC ie ALGAR—Diketones from Diacetoresorcinol-Dimethylether. 87 derivatives. The amount of the substance obtained was, however, insufficient to allow of further examination. Heywang and von Kostanecki (Ber., 35, 1902, p. 2887) condensed o-hydroxyacetophenone with oxalic ester by means of sodium, and on heating the oily diketone thus formed with alcoholic hydrochloric acid they obtained chromone. In order to prepare dichromone this reaction was tried, using diacetoresorcinol instead of o-hydroxyacetophenone. The reaction was tried under various conditions, but no indication of the formation of a diketone was obtained, diacetoresorcinol being recovered unchanged. Of the diketones described in this paper diacetylaceto-resorcinol- dimethylether and di-a-phenylacetylaceto-resorcinol-dimethylether are colour- less compounds ; dianisoylaceto-resorcinol-dimethylether is coloured slightly yellow, but dimethoxy-isophthaloyl-dipyruvie ethyl ester, in which the methyl, anisyl, and benzyl radicals of the former compounds are replaced by the more acidic carboxyl radical, has a strong yellow colour. EXPERIMENTAL Part. Dianisoylaceto-resorcinol-dimethylether. CH30 7S OCH: CH30 - CeH; - CO - CHz2- CO L) CO - CH2 - CO: CeHy - OCH; Eight grams of diacetoresorcinol-dimethylether were dissolved with warming in 84 grams of anisic methyl ester, and 3°5 grams of sodium were added. When the reaction had subsided, the mixture was heated in an oil- bath at 120-130° C. for twenty minutes, and then allowed to stand overnight. Excess of sodium was removed with moist ether and water added to dissolve the sodium salt of the diketone. The aqueous layer was separated, and carbon dioxide was passed through it. The small amount of yellow solid which was precipitated was filtered, washed with water, and crystallized from a mixture of chloroform and alcohol, and also from xylol. On analysis it gave the following results :— 0:1429 substance gave 0°3567 CO, and 0:0727 H.O corresponding to C 68:07, H 5°65 C.sH,,O; requires C 68°54, H 5°34, The substance analysed contained a small amount of ash. Dianisoylaceto-resorcinol-dimethylether crystallizes from boiling xylol in light-yellow prisms, which melt at 232-254°C. It is insoluble in cold alcohol and acetone, somewhat soluble in cold chloroform and hot alcohol, R.I.A. PROC., VOL. XXXIII., SECT. B. [0] 88 Proceedings of the Royal Irish Academy. soluble in hot chloroform, and sparingly soluble in hot benzene and xylol. Its alcoholic solution gives a brownish-red colour with ferric chloride. Owing to the small yield obtained in this condensation, it was impossible to determine whether a diflavone derivative was formed when the substance was heated with hydriodic acid. Di-a-phenylacetylaceto-resoreinol-dimethylether. CH30 “\ OCHS C;Hs; = CH: = €0'- CH. -co ee, co: CH: -€O- CH: : C;Hs Five grams of diacetoresorcinol-dimethylether were dissolved with warming in 50 c.cs. of phenyl-acetic ester, and 2 grams of sodium wire were then added. The reaction was slow at first, but on heating for a short time on a sand-bath it became very vigorous, and a brownish solid separated. The mixture was further heated for 15 minutes on the water-bath, and allowed to stand 12 hours in a stoppered fiask. Moist ether was then added, and the diketone extracted with dilute potash. A certain amount of oily matter was obtained, which did not dissolve in the ether, and was only sparingly soluble in potash. Acidification of this oil, and attempted crystallization from alcohol, gave only a very small amount of crystalline substance. The potash extract was acidified with carbon dioxide, when a small amount of yellowish solid separated. This solid was filtered, washed with water, and recrystallized several times from alcohol, being finally obtained as colourless needles, which melted at 131-132° C. On analysis it gave the following results :— 0:1499 substance gave 0-404 CO, and 0:0805 H.0 corresponding to C 73°30, H 5-96 C.,H.,O; requires C 73°36, H 5.67. Di-a-phenylacetylaceto-resoreinol-dimethylether is soluble in cold acetone and benzene, readily soluble in cold chloroform, slightly soluble in cold, and soluble in hot, alcohol. It was not obtained in sufficient quantity to test its behaviour on heating with concentrated hydriodic acid. Diacetylaceto-resorcinol-dimethylether. CHO “OCH: CHs-CO-CH2-CO\ / CO - CH: - CO - CH; A mixture of 5 grams of diacetoresorcinol-dimethylether and 50 c.cs. of acetic ester was heated to boiling, and after 1-5 gram of sodium wire was added the mixture was warmed on the water-bath for half an hour. The Aucar—Dziketones from Diacetoresorcinol-Dimethylether. 89 semi-solid brown mass which separated was allowed to stand at the ordinary temperature for twenty-four hours. It was then mixed with ether, and water was added, until a clear, dark-red aqueous layer was obtained. On separating and acidifying it with hydrochloric acid a yellow solid was precipitated which was filtered and dried. It crystallizes from chloroform in colourless prisms which melt at 116-118° C. On analysis it gave the following results :— 0:1689 substance gave 0°3891 CO. and 0:0928 H.0, corresponding to U 62°81, H 6:10, C,,;H,.O; requires C 62°74, H 5°88. Diacetylacetoresorcinol-dimethylether is sparingly soluble in carbon bisul- phide, soluble in benzene, ether, or alcohol, and very soluble in chloroform. Its solution in alcohol gives a dark-red colour with ferric chloride. An attempt was made to prepare a dichromone by heating diacetylaceto- resorcinol-dimethylether with hydriodic acid (s.g. 1°7). The reaction was tried under various conditions, but in all cases gave unsatisfactory results. When the diketone was heated for six hours with hydriodie acid, and the oily product purified, an extremely small amount of a colourless substance was isolated, which dissolved in concentrated sulphuric acid forming a colourless solution with an intense green fluorescence. It seems probable, therefore, that this product was a dichromone, but the quantity formed was too small to admit of examination. Dimethoxry-tsophthaloyl-dipyruvie ethyl ester. CHO / OCHs C2H;00C : CO - CH - CO WS CO - CHz - CO: COOC2H; 10 grams of diacetoresorcinol-dimethylether were dissolved with heating in 100 e.cs. of diethyl oxalate. The solution was cooled, and 4:2 grams of sodium wire were added. When the reaction had subsided, the mixture was warmed in an oil bath to 120° C. for fifteen minutes, and then allowed to stand in a stoppered flask for twelve hours. A brown solid separated, from which excess of sodium was removed with moist ether. Water was then added to dissolve the sodium derivative of the diketone. The alkaline solution was separated from the ether and acidified. The yellow solid which separated was filtered and crystallized from a mixture of chloroform and alcohol. On analysis it gave the following results :— 01528 substance gave 0°3179 CO, and 0:0749 H.O corresponding to C 56°74, H 5-44 CHO requires C 56°87, H 5:21. 90 Proceedings of the Royal Irish Academy. Dimethoxry-isophthaloyl-dipyruvic ethyl ester crystallizes from alcohol in yellow felted needles, which melt at 186-188°C. It is soluble in cold chloroform and acetone, insoluble in ether, sparingly soluble in benzene and alcohol, and readily soluble in boiling alcohol. Its alcoholic solution gives a brownish-red colour with ferric chloride. Concentrated sulphuric acid colours the crystals orange, and dissolves them, forming an orange solution. Heating the diketone with concentrated hydriodic acid, with a view to obtaining a dichromone derivative, gave a non-crystallizable tarry product. The substance was also heated with concentrated hydrochloric acid, in a sealed tube, at 160° C., for three hours. The product in this case was so resinified that it could not be erystallized. Dimethoxy-tsophthaloyl-dipyruvie methyl ester. CH30 ae OCHs ( C . (0) . 2° . . . HsCOOC - CO: CH2*CO\_// CO: CH2 - CO - COOCHS Dimethoxy-isophthaloyl-dipyruvie methyl ester was prepared by the conden- sation of diacetoresorcinol-dimethylether with dimethyl oxalate, in a manner similar to that described for the condensation with diethyl oxalate. On analysis it gave the following results :— 0:137 substance gave 0:2724 CO; and 0:0629 H.O corresponding to C 54:22, H 31 C\sH, i} +2 NaCl HC\ /CH HO\ /CH Cc Cc | b Na O.CO.CH3 Rearrangement with the elimination of water then takes place with the formation of te) CHs .C/\C: CH: i CH; . CO.C\_/C. CO. CHs C bx Compound M. P. 75° C. This compound on warming with sodium hydroxide changes into diaceto- orcinol as follows :— OH HCH oy) SENN CH CHs.C/\C : CH2 CH3;.C C.ONa CH3.C/\C. OH (test ea le CH;. CO.C\ /C.CO.CHs CH; .CO.C\7C.CO.CH; CH3.CO.C\ /C.CO.CH; Cc Cc Cc ! | | OH OH OH Intermediate compound, diaceto-orcinol. The only formula which can be assigned to diaceto-orcinol prepared by this method is that shown above. On application to Professor Collie he was kind enough to provide a specimen of the compound obtained by him, and this was compared with the substance prepared by the aluminium chloride method. The melting-point of the mixed substances corresponded exactly with that of the individuals, and in other respects the compounds were identical. It therefore seems certain that the two acetyl groups are in the positions indicated in Collie’s formula. This view issupported by the fact (Tambor, Ber., 39,1906, p. 4038; and 41, 1908. pp. 787 and 793) that when only one acetyl group is introduced into orcinol it preferably enters the position between the two hydroxyls, thus :— CO . CH; \ HO OH | NA CH; Auear— Unsaturated Ketones derived from Diaceto-Orcinol. 111 Whereas resorcinol under similar conditions forms resacetophenone. HO /\ 0H Ly \ /C0.CHs r The general properties of the ketones described in this communicatioi closely resemble those of similar ketones prepared from diacetoresorcinol (Ryan and O’Neill, loc. cit.); but efforts to obtain diflavone or diflavanone derivatives from them did not meet with success. When the diacetate of dibenzylidene-diaceto-orcinol was brominated, a certain amount of hydro- bromic acid gas was eliminated, and when the product was heated with alcoholic potash, a small amount of a brownish amorphous substance separated which did not crystallise, and did not show any of the properties of a diflavone derivative. Similar treatment of dibenzylidene-diacetoresorcinol readily gives colourless crystals of ditlavone. When the same reaction was tried with dianisylidene-, diveratrylidene-,and dipiperonylidene-diaceto-orcinol there seemed to be no formation of diflavone or dicoumaranone compounds. An effort was made to prepare diflavanone derivatives by allowing a mixture of diaceto-orcinol and benzaldehyde or anisaldehyde to stand for some days in the presence of alcoholic hydrochloric acid. ‘he results were, however, negative, and further attempts to prepare these compounds were hampered by the difficulty of obtaining a supply of orcinol. During the preparation of dipiperonylidene-diaceto-orcinol two isomeric substances were isolated, one of which melts at 236°-237° U. and the other at 248-249° C. The former is referred to as a-dipiperonylidene-diaceto- orcinol, and the latter as (-dipiperonylidene-diaceto-orcinol. The (3 com- pound is apparently produced from the a compound by the action of warm hydrochloric acid. The a derivative dissolved in potash, but the 8 derivative seemed to be entirely insoluble. It is, therefore, possible that the 8 compound has either of the flavanone formulae I or II. CO 5 CH: CH 5 CeH3: O2CH2 CH, CO | © ZS O HON NCR .CsH3: 020H2 CH202: CsH3. CH \ _ ZN /CH. UsHs: 020H2 O || | N/\/CH2 \/\/CH2 CH; CO CH; CO the formula of the a compound being :— Co. CH : CH 5 C.H3 5 O2CH2 HO/\, 0H ee \/CO.CH: CH. CsHs : 02CH2 CHs Owing to the small amount of the (3 compound isolated, it was impossible to further examine the substance with a view to determining its constitution. [S 2} 112 Proceedings of the loyal Irish Academy. EXPERIMENTAL PART. Diaceto-oreinol. CO . CH: HO/\0H K /oo . CH; CHs3 Orcinol diacetate was prepared by adding 20 c.cs. of acetyl chloride to 15 grams of orcinol in a round flask under a reflux condenser. When the energetic reaction which set in had subsided, the flask was heated in an oil-bath to 130° C. until the evolution of hydrochloric acid ceased. Diaceto-orcinol was obtained from the diacetate by adding 15 grams of powdered anhydrous aluminium chloride to the cold diacetate, the temperature of the mixture was then gradually raised to 160° C., and the heating continued for four hours. The molten mass was cooled, cold water added gradually, and finally some hydrochloric acid. The crystalline residue of diaceto-orcinol was filtered, washed with water, and recrystallised several times from alcohol. About 7 grams of diaceto-orcinol were obtained :— 0-1709 substance gave 0°3955 CO., and 0:0919 H,O corresponding to C 63:11, H 5:97 C,,H,,0,, requires C 63-46, H 5-76. Diaceto-orcinol crystallises from alcohol in colourless needles which melt at 93°5-94-5° C., are somewhat soluble in alcohol, and easily soluble in ether, acetone, chloroform, or benzene. An alcoholic solution of the substance gives a reddish-brown coloration with ferric chloride. Diaceto-orcinol-dimethylether. CO . CHs CH;0 /\.0CH; \ co . CHs CH; A solution of 6 grams of potash in 15 c.cs. of water was added to 5 grams of diaceto-orcinol ina round flask, The mixture was heated on the water-bath and 10c.cs. of dimethyl sulphate were added in about three instalments, the contents of the. flask being vigorously shaken during the reaction. The liquid was then cooled, made slightly alkaline with potash, and the yellowish oil which separated was extracted with ether. The ether solution was washed with dilute potash, then with water, and dried over calcium Aucgar— Unsaturated Ketones derived from Diaceto-Orcinol. 118 chloride. On evaporation of the ether, diaceto-orcinol-dimethylether remained behind as a colourless oil. It was purified by distillation in vacuo :— 0:1645 substance gave 0°3988 CO, and 01031 H,O corresponding to C 66°11, H 6:96 C,3H,,O, requires C 66:10, H 6°78. X Diaceto-orcinol-dimethylether is a colourless oil which boils at 195-197° C. under 30 mms. pressure. It is insoluble in water, but easily soluble in alcohol, acetone, ether, chloroform, or benzene. A solution of the substance in aleohol gave no coloration with ferric chloride. Dibenzylidene-diaceto-orevnol. CO. CH: CH. OsHs HO/\.0H a0 _CH: CH. GcHs CH3 About 12 c.es. of 50 per cent. sodium hydroxide were added, drop by drop, . to a solution of 4 grams of diaceto-orcinol, and 10 ces. of benzaldehyde in 100 c.es. of boiling alcohol, and the mixture heated for about 45 minutes. When the liquid was cooled no solid was precipitated, but on addition of alcoholic hydrochloric acid a yellow solid separated out. This substance was filtered, washed with water and alcohol, and purified by dissolving it in acetone and reprecipitating it with alcohol. In this manner a light yellow non-crystalline substance was obtained. Efforts to obtain it in a crystalline condition were unsuccessful :— 0:1594 substance gave 0'4567 CO, and 0:0812 H.O corresponding to C 78:13, H 5:66 C.;H2,O, requires C 78:12, H 5:21. Dibenzylidene-diaceto-orcinol is an amorphous light-yellow solid which melts about 143-153°C., and is almost insoluble in alcohol, easily soluble in ether chloroform, acetone, or benzene It dissolves in concentrated sulphuric acid, forming an orange-red solution. Ferric chloride colours its alcoholic solution brown. Dianisylidene-diaceto-orcinol. CO. CH: CH. CsHy. OCH; HO/\ 0H \ /C0.CH: CH. (cH; . OCH; CH To a solution of 4 grams of diaceto-orcinol, and 10 ¢.es. of anisaldehyde in 114 Proceedings of the Royal Irish Academy. 100 c.es. of boiling alcohol, was added, drop by drop, 12 ces. of 50 per cent. sodium hydroxide, and the liquid was kept gently boiling for about half an hour. The mixture was then cooled and acidified with alcoholic hydrochloric acid. The yellow solid which was precipitated was filtered, washed with water, then with a little alcohol, and purified by dissolving in acetone, and reprecipitating with alcohol. It was finally recrystallised from benzene. 0-175 substance gave 0-4659 CO: and 0-088 H.O corresponding to C 72:6, H 5°38 C.-H.,0; requires C 72:97, H 5:4. Dianisylidene-diaceto-orcinol erystallises from benzene in orange prisms which soften at 228°C., and melt at 231-232°C. It is almost insoluble in alcohol, and soluble in chloroform, hot benzene or hot acetone. The crystals are coloured dark red by concentrated sulphuric acid, dissolving to a brownish- red solution. An alcoholic solution of the substance gives a brown coloration with ferric chloride. ; Diveratrylidene-diaceto-orcinol. CO.CH: CH. C,H: : (OCHs)z HO/\.0H \ /CO . CH: CH. CeHs : (OCHs)2 CH; Diaceto-orcinol (4 grams) and veratric aldehyde (10 grams) were dissolved in 100 ces. of boiling alcohol, and 12 c.cs. of 50 per cent. sodium hydroxide added gradually to the solution, which was then heated for about ten minutes. The mixture was cooled and acidified with alcoholic hydrochloric acid. The yellow precipitate which separated was filtered and washed with water and alcohol. It was purified by washing with a small quantity of hot acetone and crystallised from benzene About 2 grams of the substance were obtained. 0-1629 substance gave 04106 CO, and 0-084 H.O corresponding to C 68-74, H 5-72 C.;H.:0; requires C 69-01, H 5-89. Diveratrylidene-diaceto-orcinol crystallises from benzene in orange-yellow prisms which melt at 188-189-5°C., and are almost insoluble in alcohol or acetone, but soluble in chloroform or hot benzene. Ferric chloride produces a deep-brown coloration in an alcoholic solution of the compound. Con- centrated sulphuric acid dissolves the crystals, forming a cherry-red solution. A.car— Unsaturated Ketones derived from Diaceto-Oreimol. 116 a-Diprperonylidene-diaceto-orcinol. CO .CH: CH. C,H;: 02CH2 HO/\0H | \ /CO . CH: CH. Css: 0:CH: CH; A mixture of diaceto-orcinol (5 grams), piperonal (8 grams), and absolute alcohol (70 c.es.), was heated to boiling, and 10 c.es. of 50 per cent. sodium hydroxide were added gradually to the solution. The mixture was heated for fifteen minutes, then cooled and acidified with alcoholic hydrochloric acid. The yellow precipitate was filtered, washed with water and alcohol, and crystallised from benzene. It was further purified by solution in alkali and precipitation with hydrochloric acid. The precipitate was filtered, washed with water, dried and crystallised from boiling xylene. 0163 substance gave 0-4098 CO, and 0:0698 H.O corresponding to C 68°56, H 4:75 C.;H,.O; requires C 68°57, H 4:27. a-Dipiperonylidene-diaceto-orcinol crystallises from boiling xylene in orange-yellow prisms, which soften at 234° C. and melt at 236-237°C. It is insoluble in alcohol, but soluble in chloroform, and in hot benzene or xylene. A solution of the substance in boiling alcohol gives a faint brownish-yellow coloration with ferric chloride, the weakness of the colour being probably due to the fact that the compound is very sparingly soluble. Concentrated sulphuric acid colours the crystals purple, dissolving them to a purple-red solution. [3-Dipiperonylidene-diaceto-orcinol. During the preparation of the a compound, portion of the product was heated with chloroform and a little alcoholic hydrochloric acid, with a view to decomposing any sodium salt which might be present. As a result a yellow, very insoluble substance was formed. The mixture was evaporated to dryness, and the residue crystallised from boiling xylene, in which it is much less soluble than the a compound. 0°1705 substance gave 0°4291 CO: and 0:0714 H.O corresponding to C 68°63, H 4-65 C.;H..0; requires C 68°57, H 4:27. 116 Proceedings of the Royal Irish Academy. B-Dipiperonylidene-diaceto-orcinol crystallises from boiling xylene in yellow prisms, which soften at 245°C., and melt at 248-249° C. It is much less soluble in the usual organic solvents than the a compound, and appears to be quite insoluble in potash. Its behaviour towards concentrated sulphuric acid, is similar to that of the «a compound, whilst ferric chloride produces a somewhat fainter colour than that obtained with the latter. PROCEEDINGS OF THE ROYAL IRISH ACADEMY VOLUME XXXIII SECTION C.—ARCHAOLOGY, LINGUISTIC, AND LITERATURE. DUBLIN: HODGES, FIGGIS, & CO., LTD. LONDON: WILLIAMS & NORGATE 1916-1917 THE ACADEMY desires it to be understood that they are not answerable for any opinion, representation of facts, or train of reasoning that may appear in any of the following Papers. The Authors of the several Essays are alone responsible for their contents. Dustin: Printen at THe Untversity Press ry Ponsonny anp Gruns. CONTENTS SECTION C.—ARCHAOLOGY, LINGUISTIC, AND LITERATURY. PAGE -Armstrone (HE. C. R.), F.S.A.:— On some associated Finds of Bronze Celts in Ireland. (Plates XLVI-XLVII), . ; 0 6 : 6 ; . oli Bernarp (Most Rev. J. H.), D.D. :— The Foundation of Tintern Abbey, Co. Wexford, . ; 6 Ol Biecer (F. J.) :— Some recent Archeological Discoveries in Ulster. (Plates I, II), . 1 Dix (EH. R. McClintock) :— List of Books and Tracts printed in Belfast in the Seventeenth Century, 6 : : ‘ ie j ; F 73 An early Dublin Alioansels (Plates XXX, XXXI), . : . 225 Kane (W. F. de V.), M.A. :— Additional Researches on the Black Pig’s Dyke. (Plate XLVIII),. 5389 Lawtor (fev. H. J.), D.D., Lirr.D. :— The Cathach of St. Columba. (Plates XXXIII-XXXVIII), . . 241 Macauister (R. A. §.), D.Lirr., F.S.A. :— On an Ogham Inscription recently discovered in Co. Wicklow. (Plate XXXII), : : : . : . 230 Notes on certain Irish Tee pone (Plate VN), o . 81 The History and Antiquities of Inis Cealtra. (Plates VII- _XXVII), 98 On a Runic Inscription at Killaloe Cathedral, . : : 6 . 493 Robert Downing’s History of Louth, . 0 . 499 A Report on some Excavations recently conanetedt: in Co. Galway. (Plates XLIII-XLV), . : : 6 : é : . 505 Mazarry (fev. J. P.), D.D., C.V.0. :— On the Introduction of the Ass as a beast of burden into Ireland, . 580 Murray (fev. R. H.), Lirr.D. :— Unpublished Letters of William Penn, . 5 : é : . 288 4 Contents. PAGE Rosryson (ev. J. L.), M.A. :— On the Ancient Deeds of the Parish of St. John, Dublin. (Plate XXIX), . : : : : . : ‘ c : . 175 Westroer (M. 8. D.) :— Note on Irish Money Weights and Foreign Coin current in Ireland. (Plate V). : : : . : : : : : : 43. Westroep (T. J.), M.A. :— On certain typical Earthworks and ia in Co. Limerick. (Plates ITJ-IV), . 3 On certain typical miaeenwork anil Sane. “walls in Co. Tameriok: ii—The Royal Forts in Coshlea. (Plates XXXIX-XLII), . 444 ERRATA. SECTION C. p. 65, l. 21, for 1738 read 1694. p- 100, foot-note 2, for Angciss read Misfortune. p. 134, 1. 22, omit the sacrament of. p- 242, last line, for xiv read x. p. 454, 1. 14-16, should read Eoghan. . . who having been exiled to Spain married: Beara . . . He had a son Oilioll Olom who married Sabia. . . p. 478, for (for Mr. P. J. Lynch’s survey of the dolmens) by Dr. George Fogerty read by Mr. P. J. Lynch, and delete the foot-note. PROCEEDINGS or THE ROYAL TRISH ACADEMY PAPERS READ BEFORE THE ACADEMY IT, SOME RECENT ARCHAEOLOGICAL DISCOVERIES IN ULSTER. By FRANCIS JOSEPH BIGGER, M.R.1.A. Pirates I-II. Read NovemsBer 8, 1915. Published Fepruary 14, 1916. I. A FIND OF THREE LARGE CINERARY BURIAL URNS FROM THE BANKS OF LoucH NEAGH IN ANTRIM. Puate I. THESE two perfect urns and the fragments of another were turned out last spring in a gravel pit by James Nimmo, in the townland of Creggan, in the Parish of Duneane, about half a mile from the banks of Lough Neagh, in Antrim. The whole of this district is thickly studded with earth-forts, mostly of the single-ring and fosse type with depressed centres. Many are still occupied by primitive cottages and hag-yards, showing continuous occupation for centuries. The place-names all around also prove the land to have been closely inhabited. The urns were discovered in a gravel bank which was being removed to provide road material. The bank rises gently above a small stream running into the lake to the south. There was no surface-indication of their presence, and there was no cist or stone enclosure around any of them. They were about two feet below the surface, a few feet apart, inverted, resting on rude slabs, covering the bones which were piled below them. On their discovery, word was sent to me, and I at once visited the place. The two urns had R.I.A. PROC., VOL, XXXTNJ., SECT. C, (1) 2 Proceedings of the Royal Irish Academy. been carefully removed, with the bones, to the adjoining house, and I assisted in gathering up the fragments of the one which had been unintentionally broken in the picking of the gravel. The workmen scarcely noticed the first urn, it being of similar colour to the gravel, until it fell down in pieces, smashed by their tools and the fall. This loss, however, saved the other two urns, which were abstracted with care, and the bones lifted and placed in them, together with the portions of the small bronze knife-dagger, which is of particular value as proving the age of the burial. It is illustrated (Plate I, fig. 4). The bronze object had also undergone the influence of fire, doubtless, at the time the bodies were consumed. The calcined fragments of bones were submitted by me to Dr. Scharff, of Dublin, and Professor Symington, of Belfast, and they examined same jointly, and have made a joint report, which is as follows :— “These fragments of bones have been carefully examined by Professor Symington, of Belfast, and Dr. Scharff, of Dublin, in conjunction, and they agree that some of the fragments are human, but certainly some of them are animal, probably either horse or ox. They have been burned and broken into such small fragments that it is almost impossible to identify them with certainty. Signed, “J. SYMINGTON. Rk. F. ScHarrFr. “12th June, 1915.” This is a remarkable report, because it goes to prove that the bodies of the warriors or chieftains were burned, most probably with their horses, and this in the Bronze Age. We thus have simple urn burial, in large, well-made, ornamented vessels, of the burnt fragments of human bodies, combined with animal remains, and a small bronze knife-dagger. No superficial monument was found, and there was no stone or other enclosure about the urns. They were discovered accidentally during labour ; no grave was destroyed and no ancient monument disturbed. The two complete urns are of the largest class, varied as to ornament and shape, as can be seen by the illustrations (Plate I). The fragments of the third urn have been reconstructed, and show that it was smaller than the other two, but of the same type. It may possibly have been ornamented, but the fragments are worn, and it is impossible to be certain on this point. The largest vessel measures 12} inches in height and 11} inches in diameter at the mouth; the second urn is 12 inches high and 124 inches in diameter at the mouth; and the third as retained measures 10 inches in height and 103 inches in diameter at the mouth, Breger—Some Recent Archaeological Discoveries in Ulster. 3- The two larger urns are ornamented; in one the upper portion of the vessel is decorated with a diamond pattern filled with three and four lines, and in the other with a simple herring-bone design. The urns and fragments belong to what is termed the “over-hanging rim type,” and probably may be placed in the latter portion of the Bronze Age. This type is fairly repre- sented in finds made in the counties of Antrim and Down; and Dr. Abercromby’s illustrations,! 1, Plate LX XIX, figs. 207-10, of urns found in these counties may be compared with the vessels which form the subject of the present paper. These urns have now been presented to the Academy, and are to be displayed in the National Museum as found with the actual earth and gravel around them. II. Dummy SToNE CELT FOUND IN A SEPULCHRAL CHAMBER IN THE PARISH OF CARNMONEY IN ANTRIM. In the spring of 1915 a friend of mine, a farmer named David Smith, who lives quite close to me in my own parish, was ploughing up new land on the southern slope of Carnmoney Hill,in the townland and parish of that name in the County of Antrim, when the nose of his plough turned aside a flag-stone. revealing a small chamber about 3 feet by 2 feet wide and a foot deep, built Fic. 1.—Dummy Celt from Carnmoney, County Antrim. with ordinary flat field stones. The chamber was quite empty, only a little soil and dust being in it, excepting this dummy celt. The farmer at once informed me, and I visited the place, confirming his information. The celt is at its longest 85 inches, and at its widest 3} inches and -inch thick. It is carefully worked, and the edges are well ground, but it shows no appearance of usage. It is made of shale (fig. 1). 1 Bronze Age Pottery, vol. ii. [1*] 4 Proceedings of the Royal Irish Academy. To carry on his work, the flag-stone was replaced by the farmer, and I brought away the celt. There was nothing unusual in the surroundings ; no earth-work was near, and there is no known souterrain in the neighbourhood. A carn had stood over a cromleac on the summit of the hill a short distance away. Unfortunately most of it had been destroyed by treasure- seekers before my time. It is not unusual to find implements with stone-age burials; but it is rare to find delusive ones, although such are on record. That this celt was made to serve a purpose there can be no doubt, but why so much trouble should have been taken in the making of a dummy, when a real one could as easily have been produced, is the puzzle. Of course this is quite speculative. My sole intention is to place on record the find and the surrounding circumstances. UGE, ORNAMENTED BONE Scoop FROM LECALE, IN Down. THIs bone implement was found near my castle at Ardglass last year, and given to me by the finder. Its age is quite uncertain. In the Journal of the Royal Society of Antiquaries of Ireland, vol. xlv, p. 173, five specimens are Fic. 2.—Bone Scoop from Lecale, in Down. described (two similar to this one), some of which are stated to have been found in erannogs, whilst I have one made by a man still living in Belfast, in BiccEr—Some Recent Archaeological Discoveries in Ulster. 9 the ordinary course of his occupation, from a sheep’s trotter, and used as a scoop. These were made, and sold for sixpence, and prove an heredity worth noting. The Lecale specimen, however, shows considerable age and much usage, whilst the ornamentation on the handle is elaborate (fig. 2). Its total length is five inches. It is made of the metatarsal bone of a sheep or goat. Some writers connect these implements with apple-coring. Lecale was not an apple-growing centre; but that is a small point, as such tools were easily carried any distance, and may have even come in by the port. IV. THREE WOODEN VESSELS FOUND IN A BoG NEAR DUNAMANY IN TiR EOGHAIN. THESE three vessels were dug out of a bog-cutting in the townland of ‘ir Kernahan, in the parish of Donaghkeady, in ‘I'fr Eoghain, by John Gamble. The small vessel is now much contorted ; it was about 6 inches in diameter, and 2 inches in depth. The medium-sized vessel was about 7 inches in diameter, and 2} inches in depth. Both these vessels were circular. Fig. 3.—Wooden Vessels from Dunamany, in Tir Eoghaia. The greatest interest centres around the large oval vessel on account of its skilful carving, with its four legs, out of one piece of wood. This is quite unusual, proving the use of excellent tools in its manufacture. It was well shaped, cleanly cut out, and neatly rounded. he feet have been carved out of the solid block of wood, and equally spaced on the base. There is no 6 Proceedings of the Royal Irish Academy. appearance of handles, and any tool markings that can be traced were done with fine implements. All are illustrated on a much-reduced scale in fig. 3. The age of these vessels is quite problematical; but judging by their con- dition, and the situation of their discovery, they are derived from a crannog of medieval date. Professor A. Henry has kindly examined the vessels, and reports that the large vessel with four legs is made out of a piece of either sallagh or poplar wood, the vessel next in size being made from a picee of alder wood. We EARLY BRONZE PROCESSIONAL CROSS FROM DONAGH, IN THE DIOcESE OF CLOGHER, ALSO THE ANCIENT HIGH Cross OF DONAGH. PuatTeE II. I cau this rare specimen of an Irish processional cross the bronze cross of Donagh, because it was found there. In 1911 I was stopping at Glasslough, with my friend Mr. Shane Leslie, and after perusing Shirley’s “ History of County Monaghan ” (p. 295), we went to the old graveyard of Donagh, on the summit of a hill, to look for the cross mentioned by Shirley. I was assured on all hands that it had disappeared, and so it had. A close examination of the church ruins convinced me, and I soon satisfied my friend that the old church had been used by the Planters. My friend was of the opinion that the present Protestant church was on the site of the Planters’ church. Having got so far, we looked around to find any trace of the cross, thinking it might have been destroyed or removed by the older race to preserve it from desecration, as Monaghan had much turbulence in the Plantation and even later times. After diligent search I came on a mossy stone level with the ground. On removing some grass and earth, I found it unshakable, thus proving it had some depth in the ground. Further excavation revealed the head of the lost cross of Donagh. We did not delay in having it unearthed and set up on its old site in a firm foundation, so that now it is visible for a long distance in every direction. It stands about 5 feet high, and 3 feet wide in the arms, and has upon its east face the figure of our Lord carved in the old Irish way. One thing leads to another. Mr. Shane Leslie still searching a short time afterwards, discovered in a disused room above the sacristy of the old Catholic church at Donagh this bronze crucifix. He at once consulted with me regarding its preservation. My advice coincided Biager-—Some Recent Archaeological Discoveries in Ulster. 7 with his wishes. It should be conserved as far as possible without any alteration, save replacements, in the condition in which it was found, and restored to its original use and benefit. This has been done in a way I will tell you, and with the result you can see. It is now in the care and trust of the Most Rev. Patrick Mac Kenna, Bishop of Clogher, and is kept at Saint MacArtin’s Cathedral, in Monaghan. ‘his we considered a more worthy way of treating so valuable a relic than even having it carefully preserved amongst the Irish Antiquities in the Royal Irish Academy’s Collection in the National Museum, or running the risk of its retention in a private collection. When found it had only two beads, the one at the right side of the head, and the one at the foot; and the top jewel was missing. The bead at the left side of the head is one I was fortunately able to supply, as well as the jewel at the head of the cross. This bead is an ancient one, and the jewel is most appropriately a polished pebble from Cranfield on Lough Neagh. The cross itself is made of two lengths of brass fastened at the centre one 14 inches long, and the cross-bar 9 inches long, both about # inch wide and } inch thick. ‘he head and ends of the arms enlarge into circles to hold the jewels. The figure from the drapery and head-dress appears to be of the thirteenth century. It corresponds in some respects to the figure on the cross of Tuam, which is of that date. It is much worn by use. The feet are crossed, and with the hands are fastened, with bronze nails, right through the shaft and arms of the cross. The head inclines to the right ; the figure is not what might be called a hanging figure of the dead Christ, but has some of the Celtic spirit of life, with the arms widely expanded, and the eyes open. The remarkable feature is the addition of the beads. They are fastened with bronze pins through the cross, and are clearly of much older origin than the crucifix. The original blue enamelled bead at the right side is a particularly fine specimen, with circular ornament upon it in red, purple, and white. The original bead below the feet is of purple glass; the ancient added bead to the left is of white glass. Further along the arms of the cross are two other openings which are also doubtless made for nails to fasten on beads or other ornaments. There is also one at the back of the head. The original jewel at the right hand is of crystal; that to the left is of purple glass. The one at the head is the Lough Neagh pebble. The jewels on the arms are simple, clasped with bronze indented hoops fastened to the shaft, whilst the one at the head has a similar clasp with a twisted wire around it, and a circle of ball ornament, giving it the appearance of an earlier date than other features would warrant. This ornament may have originally existed around the side jewels, and been 8 Proceedings of the Royal Irish Academy. destroyed. The back of the crucifix is perfectly plain. The base of the cross-shaft is slightly expanded with three nail holes in it, clearly showing that it had been fastened into a base or shaft to be held erect. These three holes were used in the new shaft which has been added, as I had no doubt this crucifix was the terminal of a processional cross. A word for the shaft as now restored. Every piece of it is Irish work and material. of ancient Irish timber from the oak woods of Ballinderry, in Co. Antrim, used for centuries in the old O’Neill church at Portmore. The serpent head fastening the shaft to the cross is copied from that of the cross of Cong, interlaced with four jewels, marbles, red from Munster, green from Connacht, black from Leinster, and white from Ulster. There is a Celtic bronze ferrule, and an interlaced bronze cross upon the bulb in the centre of the shaft with It is the simple inscription :— Tus. Sean. ua Laplars on Cpoip peo. Do. Naoth. Macdsitain. A. 1911. C. PLATE I. SECLALG. XXXIIL., Proc. R. I. ACAD., VOL. 4 ~ ‘ cpa tn ish | il Adit at Hh i MT fi yf aa 4 iy i) Ma fT). ra He HAA WP Wl) "h) Hy sy Mh YP mA ie Wy um b3 mi Missy i Mig i LH Uy) Yi ea =sS ee ee = : _Cinerary Urns found at Creggan, County Antrim, and Bronze Dagger found in one of them BIGGER.—ARCHAEOLOGICAL DISCOVERIES IN ULSTER. SAN, ‘fe ‘MHLST(Q] NI SAINAAOOSIG IVOINOTIOUVHOAY— AAVO I “BS0H MT 'V Aq “o704q s ‘ yseuog Jo ssoag ystyy ay.—'1 “SIT “IDYSO[D JO ssa001q] ay} ur ‘YsvuoG JO xylonID [vUOIssad.01g ezuoIg oY [.—"?e “DI *IOYBOTD JO esad01qy ay} ur ‘II FLv1g ‘9 Lots “TIIXXX “IOA “avoVy [YU 90%d 0 —— = (I Ko) (L— oe ON CERTAIN TYPICAL EARTHWORKS AND RING-WALLS IN COUNTY LIMERICK. By THOMAS JOHNSON WESTROPP. PuLates III anp IV. Read Novemper 30, 1915. Published Marcu 16, 1916. THE study of early Irish forts on any kind of scientific system was hardly known in Ireland a quarter of a century ago. This may seem a startling misstatement, till we see that isolated general descriptions (usually accompanied by some universal theory, and rarely even approximately full and accurate) alone recorded the few forts described in accessible publica- tions. In theory a baseless fabric (supposed to be vouched by a poem written 1000 years after the supposed date of the events it records), as to the stone forts being all the work of a little tribe of Firbolgs, alone held the field—the earth forts, of course, were “Danish.” A few people remembered that the originator of the Firbolg theory, O'Donovan, had the sound common sense to make many exceptions to it, but most persons gave it a universal applica- tion. Generalization was impossible; some twenty stone forts, two or three promontory forts, and a small number of earthworks alone were described, some most incorrectly ; hardly a correct plan had been published. The great work of Lord Dunraven only describes two types of stone forts, and does not give a single accurate plan. For earthworks, a foreign antiquary could find little but the old-fashioned, and not always correct, views and plans of Wright’s Louthiana and a few plans of earthworks, like Tara and Usnach, in the publications of this Academy and of the Royal Society of Antiquaries, under its various titles. The only efforts to utilize comparative archaeology were vague comparisons of our ring-walls with the city forts of Etruria and Greece. Rarely do we find even a hint in Irish papers on the forts that similar structures remained in Great Britain ; none of our antiquaries strove to study what was being done on the Continent. A large class regarded round towers, early churches, and stone forts as the peculiar property and glory of Ireland, and resented any attempt to describe the forts of other countries. Much of the unprogressive character of Irish archaeology springs from its exponents taking no trouble to keep up to date (by reading the R.I-A. PROC., VOL. XXXIII., SECT. C. [2] 10 Proceedings of the Royal Irish Academy. proceedings of the societies), and confining themselves to repeating the many times refuted theories, identifications, and catchwords in books half a century behind modern archaeology. People quoted with approval the poem of the race, “taller than Roman spears,’ making “ their mystic forts,” but were too careless to look in old Irish literature and in European archaeology for what would have stripped the epithet “mystic” from the common, practical home- steads of all early and mediaeval periods of Irish peoples. Since 1890 we have had to collect our material from every source—the library, and still more the field, before we had any reasonable amount of matter for generalization ; let this be an excuse for us workers where the inferiority of our general views is criticized. Yet much has been done in Connacht and Munster, less in the other provinces. In Munster much material is available in Clare and Kerry, and round certain districts in Cork, Waterford, and Tipperary.’ Only one county has been too much neglected-—Co. Limerick. Still, as ever, a pioneer and beginner, I crave the forbearance of the Academy for laying before it an instalment of this necessary work, covering, it is true, most of (if not all) the types of forts, but giving far fewer examples of each than in previous papers on the other counties of the province of Munster. Some make little allowance for an imperfect survey, but hardly anything else is possible in Ireland, and it is most helpful to publish even such a paper. “The best that every man knows dies with him,” but some- thing may be saved. The few who take interest in archaeology (as apart from historical, architectural, and linguistic questions connected with it) will perhaps judge hardly a dry survey, not even pretending to give every fort of importance. Others, more interested in weapons and implements, or in ecclesiastical and genealogical questions, may be even more like those of old, who took no interest in the fort-makers: “Ye also made a ditch between two walls for the water of the old pool: but ye have not looked unto the maker thereof, neither had respect unto him that fashioned it long ago.” The inhabitants of each district could make such a survey, but how many persons in Ireland have completed the recording of forts in even one parish ? One living far away, and only getting a week now and again to explore a district, may get much; but he certainly can never make a “ field-to-field survey,” as should be done for a parish or county, before even this limited requirement of completeness is fulfilled. Not by “flying over the country” on a motor, but by going on foot over hills and fields and wildernesses, can such work be done. British antiquaries (foreign ones seem more sympathetic) 1See an excellent paper on the earthworks round Tipperary, in North Munster Archaeological Soc. Journal, vol. iii, p. 5, by Mr. Paul Flynn. Wrsrropp—Larthworks and Ring-Walls in Co. Limerick. 11 wonder why we Irish cannot complete the record of even a few parishes. An English parish (with perhaps one or two hill forts, a church and castle, and a few lesser monuments) is finite, and can be completed; but the antiquities in the wilds, and even in tamer parts, of Ireland approach the infinite, and the longer one works the more lesser remains seem left to be recorded after that work is published. Save for the mote of Shanid, I do not recall a single earthwork in this county which has been described with any clearness before this century, and none with any fullness. I long avoided the subject myself, though my knowledge of several of the forts extends over forty years. I hoped the North Munster Archaeological Society might have done the work; but the Society did nothing of the sort, and it now (since the death of its Secretary, Dr. George J. Fogerty) seems unlikely to undertake it. I know of no local antiquary who is doing anything in that direction. So far IT have only published a general plan of the earthworks at Shanid, and detailed ones of those at Kilbradran and Cloncagh.’ Mr. P. J. Lynch had published one (with sections) of the fine mote at Kilfinnan, along with Dr. G. Fogerty’s photograph ;’ he has also given us a valuable monograph on the dolmens of the county 3 but this does not affect our subject. No elaborate paper on any fort in the county has been published as yet. Even the more casual mention is very rare. FitzGerald’s History contains a description of what are stated erroneously to be the remains on Knockfennell* (but probably on another fortified hill), beside Lough Gur; casual mention of “Danish forts” by him and Lewis, and short notes on Shanid, complete the bibliography. Of accessible written material Windele’s manuscripts have a few brief notes and rough sketch sections; while the Ordnance Survey Letters—so helpful for the churches and castles of Co. Limerick—hardly describe, even briefly, any of the forts. Irish archaeological nomenclature is in the making; old-fashioned people resent any desertion from the old but inaccurate terms —for example, from our misleading use of “cromlech”; so we use “castle” for little residential peel towers, hardly ever military works, and “abbeys” for small cells and collegiate churches. We are justified by the usage of mediaeval writers and the local nomenclature, whether in France or Ireland, in using the word “mote” for any earthwork, but will confine it here to ? Journal R. 8. Antt. Ir., vol. xliii, p. 251 (Cloncagh) ; Proc. R. I. Acad., vol. XXiv, p. 275 (Kilbradran), and vol. xxv, pl. xviii (Shanid). * Journal R. 8. Antt. Ir., vol. xli, p. 389, Kilfinnane. [2*] 12 Proceedings of the Royal Irish Academy. platform forts of any height. It no more connotates a work of one type, or of any one period, than do the cognate words “ dun,” “liss,” or “rath.” It might have been best to put together all the examples of a particular type under that head; but I think (as this paper is intended to subserve a topographical purpose as well as an archaeological one) that it is better to select in each case a district, and give all the more instructive forts in it, along with the early legends and history, so far as bearing on the forts. I will endeavour to select these groups so as to illustrate the main sections of the county. The tract from the Shannon to Dunganville may serve to illustrate the forts of Ui Chonaill, or Connello; those from the River Deel to Lough Gur may represent the north-eastern and central parts; and Bruree and the group irom Aherloe to Bruree and Ardpatrick, in Coshlea, along the Ballyhoura Mountains, are sufficient to show the earliest legends and residences of the royal Dalcassian race. Earty Divisions. As so often, the first glimmer of light shows merely the names of obscure, otherwise forgotten, tribes, such as the Gann, and possibly the Siol Gengann (the Ganganoi of Ptolemy* in 4-p. 160), then the ubiquitous Ui Catbar and Ui Corra, who had settlements down the west side of Co. Clare and Connacht, and south of the Shannon, are said to have had colonies in western Limerick. The more definite Gebtini gave their name to Askeaton (Has Geibhthine) and its island fort Iniskefty, or Inis Geibhthine. A reputed tribe of the Mac Umoir, or Huamorian Firbolgs, named Asal, is said to have given its name to Drom Asail, or Tory Hill* The Mairtinigh lay round Emly; these, and possibly kindred tribes (the Dilraighe, Margraighe, 1 Compare local Irish usage with the “ buttes” and ‘‘motes” in France: low forts as well as high are so called. ? Petrie identified the Magolicon of Ptolemy with Caherguillamore, ‘“‘an extensive city”’ (‘* Military Architecture,” mss. R. I. Acad., p. 77). If St. Mochealloc be a real person, it can hardly be Kilmochealloc or Kilmallock. Some take Brughrigh (Bruree} to be Rigia Hetera, but others place it at Athenry. An alleged poem of St. Columba (Ossianic Soc., vol. iv, p. 252): ‘To Gann ..... was given the country to the pass of Conglas (Co. Cork), and thence to the Luimneach.” From the Dal gCais territory only reaching to Carnarry it is possible that the Tuath Luimneach were once strong enough, perhaps with aid from Connacht, to hold them back. The poem gives the country from the Luimneach to Eas Ruadh (Asseroe, Ballyshannon) to Seangan and Geanand. > Revue Celtique, vol. xv (1894), p. 481. +The fleet of Luimneach plundered the Martini of Mumhan,” Book of Leinster. Wars of the Gaedhil with the Gaill (ed. Todd), pp. 15 and 227, and note, p. xlii. The Martimi of Imleach are named in O’Huidhrin’s Topographical Poem (ante 1426), line 601, but O’Huidhrin is rarely up to date, and ignores the English occupation and other changes. Todd, quoting from Book of Lismore, f. 172, and O’Curry, “‘ Battle of Magh Lena,” p. 76n, notes this tribe, and the former suggests that Cluain Comairde (Colman’s Well) was in their territory. Wesrropp—Larthworks and Ring Walls in Co. Limerick. 13 Sibenraighe, Calraighe') met at “the hill of the Banshee Aine” at Knockaney. Only the Mairtinigh appear in the Annals, which record their extermination, possibly along the Shannon, by the Northmen in a.p. 845. An important tribe, the Uaithne, called by the Normans Wethney and Owney, extended across the Silvermine Hills in Owneybeg Barony in. north-west Co. Limerick, and in Wetheney Tire, or Owney and Arra, in Co. Tipperary. The Tipperary families had migrated from Uaithne Cliach, expelled, say the legends, by the race of Cathaoir Mor. The Muscraidhe Chuire lay in Clanwilliam, round Kilpeacon, while close to them, round Crecora, were the O Colochur.’ O’Donovan places the Ara Cliach in “ western” Co. Limerick by an over- sight; he evidently meant “eastern,” and they seem to have migrated into Arra, like the Uaithne. They were Rudrician in descent, and the “Tripartite Life” places them to the east of the Ui Fidgeintiin the land of the Ui Cuanach, now Coonagh; the Saimer or Morningstar River® divided these tribes. The Aradha were raided (according to tradition) by the warriors of Eoghan, son of Ailill Olam, and are represented in modern times by the family of Ui Ciarmhaic or Kirby, of Eoghanacht Aine.‘ There were three non- tributary races, evidently later conquerors of “pre-Milesian” tribes—the Aine Cliach, the Ui Fidgeinte, and the Dal gCais. The latter were the kings of the district, under the Provincial King of Cashel, and sat “next his shoulder” at banquets; they led the van in his wars, and covered his retirement or retreat. From the tenth century they alleged an alternative right of succession to Cashel; but during the early historic period there is little or no evidence of their having obtained this position until Mathgamhain, son of Cenedig, was made king. But, in the end, the alleged alternative succession, as attributed to the will of the legendary Cathaoir Mor, was accepted, even by their opponents. Later writers interpolated Aedh of Cragliath (cirea a.p. 573) and Lorcan son of Lachtna (cirea A.D. 860-900) into reigns of the Cashel line, but there was (at best) confusion with Aedh of Cashel, whose kingship is more than doubtful? and 1 Hgerton, ms. 92, f. 37 B. 2 Black Book of Limerick, p. 100 (in 1299). This pretty name (which FitzGerald, History of Limerick, vol. i, p. 320, calls ‘‘ The River Dawn”) arose from a change of Samhair or Samer to Caimher, ‘‘the daybreak” or «‘Morning Star.” The real name seems akin to the Persian word Shamar—‘‘ The River,’’ (as the Samur, Samara, &c.), in Irish possibly “‘the Sparkler” (Joyce, ‘‘ Irish Names of Places,’ ser. 1, chap. xxvi). It is named in Prince John’s Charter to the Abbey de Magio (cr Monasternenagh), in 1185 (Cart. Roll u, John in 12, confirms it), “To Cillnarath as the Samir runs from it,’ The corrupt name occurs in 1655 as ‘the Cavoyer’”’ (Civil Survey, vol. xxxi, p. 1). +1124, ‘‘ Ui Ciarmhaic of Ane,’’ Ann. Ulster. ® He was ignored by the historians of the Ui Hacach, or O’Mahony clan, in south-west Cork. Not appearing in the most reliable lists. 14 Proceedings of the Royal Irish Academy. who was certainly a different person from Aedh of Cragliath, both bemg named by St. Brendan of Birr' as his personal friends. Lorcan was really a son of Culigan, and was not Lorean of the Cragliath line. The Dal gCais do not appear in the historic Annals before the reign of Cenedig, father of Mathgamhain and Brian Boroimhe. They appear to have split into two lines about 571, one reigning at Bruree and Singland, the other, at first more obscure, at Cragliath, near Killaloe. The first disappears after a great Norse raid in the ninth century ;? the other, by a strange chance of fortune and their own fine qualities, fought till they overthrew the Danish rule, and became kings, first of Thomond, then of Cashel, then of Ireland. Innumer- able O’Briens, MacMahons, Kennedys, and others represent them all over the world to this day. The Ui Fidgeinte claimed descent from Daire Cearbha, father of Crimthann mac Fidhe (traditional king of Ireland, but probably only of Cashel), on the borderland of history. The group was extensive; its chiefs were later known as O'Donovan; the chief branch was the Ui Chonaill of Connello (the Ui Chonaill Gabhra, or Ui Gabhra), and the later families of Ui Coilean (Collins), Ui Cinfhaelaidh (IKinealy), Ui Flan- nabhra (Flannery), and MacInneirghe (MacEnery), spring from this stem. In the time of St. Ita, the Lady of Killeedy (died a.p. 569), they were under the spiritual rule of St. Senan’s Island-Monastery of Iniscatha. Of some other divisions must be noted the Corcoithe or Gortcoyth (Macassa), of Newcastle; the Fir Tamnaighe of Mahoonagh;? the Corca Muicheat of Corcomohide, and the Ui Baithin (O’Meehans), near Ardagh; the Deisi beg lay at Knock Aine, and Cliu Mail mic Ugaine was to the south of it. Aes Greine was called Est Grene by the Normans; at the time of its greatest extension it was held by the Ui Conaing, or O’Gunnings, and extended from the Maigue up to at least Castleconnell, which fortress, and that of Carrigogunnell, once bore the tribe name as Cashlan Ui Chonaing in 1174; Castro Coning, 1242; and Carraic Ui Conaing, 1209.‘ The district included Singland, the residence of the Dalcassian King, Carthann, and where St. Patrick is said to have baptized him and his infant son, Eochaidh Bailldearg. The Caenraighe (whose name survives in Kenry, and whose land lay along the Shannon between the rivers Deel and Maigue) were said to be of the kindred of the Ui Fidgeinti, but may have belonged to the group of 1 Poem of St. Brenann of Birrha in 571, ‘‘ both are my friends.” 2 Cirea 830, before the rally of the Ui Chonaill in 839. 3 Must I again point out that Mahoonagh is Magh Tamnaige, Motawny, and Tawnagh in all authentic documents, while Magh Gamnach exists only in O’Donovan’s imagination ? Medhonach is another ‘‘ shanachee rendering.” See Proc. R. I. Acad., xxvi (c), p. 254, and Journal R.S. Antt. Ir., vol. xii (1871), p. 629. ' Carrig Gunning, 1580: Hardiman’s Map, T.C.D., No. 63. Wesrropp—Larthworks and Ring- Walls in Co. Limerick. 15 tribes, kindred perhaps to the Mairtinigh, who were reputedly “Firbolgian,” and, at another time, were under the tutelage and apparently connected with the Ui Cairbe Aobhdha. Like the Tradraighe, across the Shannon, opposite to the Caenraighe, but in the present Co. Clare, they probably claimed Telationship with different tribes in order to secure protection. They were probably weakened by the vaid of King Fiachra of Connacht at the close of the fourth century, and that formidable monarch got what proved to be his death-wound in their territory. The most prominent of their families were the Ui Maolchallain (now Mulholland) and the Ui Rosa or Ui Beagha, whose name appears at Beagh Castle and Iveross parish, where the Deel meets the Shannon. The intrusive and traditionally fugitive Deisi occupied Deisbeg or Small County. The later baronies arose (besides the ancient tribal districts represented by Owneybeg, Coonagh, Kenry, and the Connello group), Pubblebrian and Clanwilliam out of the Tuath Luimneach (O gConaing territory), Small County out of Deisbeg, Coshlea out of Fontymehyll, and Atharlach. Coshmagh is a patchwork formed out of portions of several of the divisions. Pubblebrian and Clanwilliam recall the fourteenth and twelfth century settlements of the O’Briens and Burkes. I need say no more of the Dal gCais princes, or those of the O’Donovans, till I describe the fort groups at Bruree and in Coshlea. It is only important to recall that the first under the successive princes, Lughaidh Meann, Connall Eachluath (A.D. 377), and Eanna Airigthech (after 400), conquered the central part of the present Co. Clare from the J.uimneach (or Shannon Estuary) up to Inchiquin Lake and along the hills of Burren to Luchid heath. This battle, the site of the decisive victory gained by the first king, extended his realm from Cahernarry (Carn Fearadhaig) to the present border of Clare and Galway. EARLY History oF THE Forts. When we bring together everything definite from the Annals, and even from later works, like the early ninth-century Dalcassian tracts and the Lives of the Saints, we feel that a deep gloom broods over the early history of Limerick down to the ninth century. In this period numbers of the forts were made and repaired without, apparently, any record being kept. It is, of course, more than probable that the Norse and Danish raids destroyed valuable monastic annals at MKilleedy, Mungret, Iniscatha, Cloncoraha, Knockpatrick, and possibly other monasteries; but there must be other reasons, as scattered entries occur from the middle of the seventh century. Let us note a few of these stretches of light and darkness. The Tripartite Life of St. Patrick, telling of his journey in east and north Co. Limerick in 16 Proceedings of the Royal Irish Academy. the first half of the fifth century, mentions the forts of Rathcorbry, Rath Broccain, a cathair at Mungaret, Sangal' (Singland), and Dunoacfene, near Donaghmore. Whatever doubt may he on the missionary travels of St. Patrick in Munster, the record is at least of minute topographical accuracy, and. in its relation to Co. Clare, gives an early state of affairs. At least such sections as the conversion of Carthann at Singland seem accurate in chronology and probable, for (as I said) the low-water mark of the Dal gCais territory, during the encroachments of Connacht, lay at Carnarry (Carn Fhearadaigh),? while the conquests across the Shannon made it almost necessary to the conquering princes to move up from their forts at Dunclaire, Duntrileague, and Brughrigh, nearer to the centre of their extended realm. Again, whoever wrote the account of St. Patrick blessing Corcavaskin from Finnine to the west of Donaghpatrick, visualized the view from the summit of Knockpatrick northward in the blessing® There is no trace of late tribal or monastic arrangements or attempts to exalt the Dal gCais, or make St. Patrick convert Limerick (such as the late writers made him do for Dublin), by giving them any special attention from the saint. Also Brecan or Bresail, the son of the then baptized HEochaidh, was an enthusiastic missionary north of the Shannon in the later Thomond and Aran, about A.D. 480, with no hint that his parents were pagans or semi-pagans, such as we find elsewhere told of other early saints. The Tripartite Life also falls in with that lapse in the Dalcassian princes (from Eanna, circa 400, to Dioma, circa A.D. 630) in which their territory, at least in central Co. Clare, was so little under control that at last it was ruled by a petty prince, Forannan (of a junior branch), and a mere creature of Guaire, the King of Ui Fiachrach Aidhne.* We seem to have an authentic contemporary stanza° on the two Aedhs of Cashel and Cragliath (in 571), brief, but very instructive as to the breaking away of the “ Killaloe line” from the older territory—a move attributed by the later writers to the election of Aedh to the kingship of Cashel, which is not recorded in any reliable early source. It is noteworthy that no i Was it one of the ‘hillocks” (“‘ Chnoccanaib Saingil”’) on which O’Rourke’s head was displayed in 1088? (Fragment of Tigernach, Revue Celtique, vol. xviii, p. 9.) 2 O'Donovan places it on the south border, where some person has marked its imaginary site on the Ordnance Survey maps (which are full of such imagination in this region). It was on the other border, and Carnarry was originally Carn Fearadhaig till about 1530. See the De Burgo Rental, &c. 3 He points to the Burren Hills and Echtghe, and to ‘‘yon green island in the mouth of the sea”’ (Iniscatha). All are just visible from the church site, none from any spot near Donaghmore, where some local and other writers have located the place of blessing. * See Life of St. Mochulla ; cf. Tract on the Dal gCais. 5 Cited before. Itis attributed to St. Brendan of Birrha in 371. Wesrropp— Earthworks and Ring-Wails in Co. Limerick. 17 Dalcassian king seems to be named after Eochaidh (434) till after 600. The only Saint’s “ Life” connected with Co. Limerick—that of St. Ita—has early features. The Life of St. Senan (certainly based on authentic very early material) tells us that Iniscatha and the Shannon Islands were ruled by MacTail (King of Hui Fidgente, and the Islands of the Luimneach) and another Prince, Nechtan “ Long-head,” of the same tribe. In the “ Life of St. Ita,” the importance of the Hui Chonaill (Ui Fidgeinte) and their extension to the Maigue, which runs out at the bounds of the Tuath Luimneach, tallies with the earlier Life The Life of St. Maidoe has an interesting account of a brief visit of the saint to our district, which we must study later. In the actual Annals, the Ui Fidgeinte appear from a.p. 646 (perhaps their great victory over the Norse at Shanid preserved their Annals) and Brurigh in 715. Mungret Abbey is named only in 845,’ the Luimneach, in 851, and Cluain Comharda (Cloncoraha or Colmanswell) in 844. Nota single King of the Bruree line of the Dal gCais is named, nor any of the Killaloe line till King Cenedid, in the middle of the tenth century. The tale of the last effort of the kings of Connacht to retake, not only the present Co. Clare, but evidently to sweep the Dal gCais out of their own territory, and its frustration by King Dioma, about 620-640, is not in the Annals, but only vaguely recoverable in the material collected by King Cormac mac Cuilenan of Cashel at the close of the ninth century.? This shows how little survived the great Norse raid of about 830, save legends of prehistoric kings and dry tribal pedigrees.* The literature relating to the forts, however, is better than we might expect. In the Tripartite Life in 454 forts at Mungret and Singland and Dun Bleise (Doon, in Coonagh, “ Dunleisg ” in 1559), where Fintan founded a cell about A.D. 580, are implied. In prehistoric legend Dungrot, Dun Claire, Duntri- league, and Dun Kochair Maige, or Bruree, occur.’ The fort at Mungret is called a cathair; probably (as so often) it was given to the founder of the church in its garth, in this case St. Neassan (mid-fifth century); Rath 1 Neither Life alludes to the importance of the Dal gCais, which favours the antiquity of their material. 2Tt has records back to the fifth century, if the Tripartite Life, &c., are reliable. 3 Perhaps the most serious loss to local (if not to Irish) history and archaeology is that of the Psaltair of Cashel. It seems to have been extant in the early eighteenth century, and was in Trinity College Library in 1726 (Introduction to Keating’s History) ; later citations may be from extracts. Much, however, may be extant, but its integrity is always doubtful. *A good idea of those tribal pedigrees is obtainable from the Tract on the Dalg Cais in the Book of Ballymote, ed. by the late Mr. R. Twigge, F.S.A., in the North Munster Archaeol. Soe. > Accounts of these will appear in a later section of this paper. R,I.A. PROC,. VOL. XXXIII., SECT, C, [3] 4 18 Proceedings of the Royal Irish Academy. Ui Dauaidh is named by Annals of Innisfallen in a.D. 596, Ratheguala (if it be Rathkeale) was taken about 620.’ The battle of Carn Feradaig (Carnarry), 621; also Cathair Chinnchonn of Rockbarton in 637; Aine (the Sid Eoghabhail of the Agallamh, and therefore a fairy fort, and perhaps a temple mound) in 666;? Senati or Shanid fort in 834; Cathair Cuan in the Ui Fidgeinte territory! was plundered by Brian Boroimhe in 973, while his brother King Mathgamhain was betrayed by Donnabhan mac Cathail, Prince of the Ui Fidgeinte, at Dun Gaifi in his house, believed to have been at Brurigh. Brian repaired or made forts (1002-1012) at Dun Cliath (? Aine), Dungrot and Cennabrat (near the Ballyhoura pass, Bealach Feabrat) and those on the islands of Loch Cend, Loch Gair (Gur), and Loch Saiglend, besides the important fort of Bruree or Dun Eochair Maige. Dun tri liag burned by Torlough O’Brien on his way to Thomond from Emly in 1054. Dun Aiched (an unknown fort identified by O'Donovan with Dunkip, but only from its sound), Brughrigh, and the Loch Gur forts plundered in 1086,° while the fort at Cromadh (Croom, perhaps the fine fort on the Maigue bank in Islandmore) was burned in 1149. THE FORTS IN LEABHAR NA GCEART.—The document, whatever its period (whether early or late tenth century or earlier), which gives most about forts in Munster, is the list of those claimed by the King of Cashel in the Book of Rights.© It is unfortunate that they are not given in regular order (as may be seen by the identified names), and that they rarely have any note or epithet to help us. Having recently been dealing with the districts covered by the Tain bo Flidhais, and enjoyed its vivid and most reliable topographical help, I feel all the more the want of a really helpful document for Co. Limerick. I at once avow inability to elucidate all the lst of forts, but it calls for study in this connexion, and must be faced. O’Donovan makes most of his uncertain identifications by mere guess, and curiously expends many notes in 1Ann. Ulster, 622; Chron. Scot. 623; Ann. Inisf., 616; Ann. Four M., 618; ‘**Expugnatio Rath Guala,” by Fiacha: “ well known is the strength of its beams.” 23rd Fragment of Tighernach, Rey. Celt., vol. xxiii (1902). The battle of Carn Fearadaigh in Cliu, where Failbe Flan was victor. 3 Cathair Chinn Chonn, battle won by the Munster men under Oengus Liathain (Annals Ulst., 639 or 637). + Aine, Ann. Ulster and Four Masters, 666; Chron. Scot., 663; between the Aradha and Hui Fidgente. It was a sidh or fairy mound—Sith Cliath, Knockaine (Book of Fermoy, R. I. Acad. Irish Texts, p. 9). Cuan King of Ui Fidgeinte slain in 642 at Carn Chonaill. 5 Or 1084, Chron. Scotorum. ® The list (I use the poem as probably older than the prose version) is connected with the legendary Hric of Fearghus Scannal, which, however, only applied to Ossory (Book of Lecan, f. 225 b and 229b). It is attributed to Benean, disciple of Patrick, in the fifth century, but is probably several centuries later. Westropp—Larthworks and Ring-Walls in Co. Limerick. 19 telling us that certain forts are “ unidentified ” or “ unknown to the author.” For example, “ Cathair Chuire close to the sea ” in the document is identified with Caher Gel, near Cahirciveen, though scores of Caher-names remain all round the coast. I shall give the identifications, with little comment where I am able to suggest them. Brughrigh (Bruree, Limerick) ; Muilchead Seanchua (Shanahoe near Muskerry, Cork), RosRaeda ; Cluain Uamha (Cloyne, Cork) ; Cathair Chnuis (?Caherrush on the Clare coast) ;+ Cathair Fhionnahhrach (a stone fort near Kilfenora, Cil Fhionnabhrach, Clare, perhaps Ballykin- varga*); Cathair Thuaighe ; Cathair Ghleanna Amhnach (Glanworth, Cork), Cathair Chinn Chon’ (Rockbarton, Limerick) ; Dun fir Aen Cholga (? Dunae- nir, Kerry); Dun Gair (Lough Gur, Limerick) ; Cathair Meathais ; Teamhair Shubha (?Teamhair Luachra); Air Bile (Ardvilly near Ballymacelligot, Kerry), “the great wealthy red”; Aenach mBearrain (Lissrawer, Burrane, Clare®); Magh Caille; Ard Chonaill “the meeting place of hosts,” Ard mic Chonaill’ with Ard Ruidhe (Caherconnell, Cahermacconoal and some fort on the hills above them, Carran, Clare); Tuaisceart Muighe; Magh Saire, “worthy of reckoning with the three Aras of the sea” (Aran Isles, then part of Clare); Aenach Cairpre (Manisternenagh, Limerick) ;’ Druim Mor; Druim Caein. (These two names very common. O’Donovan thinks the first Dromore near Mallow, Cork. Perhaps the last is Dromkeen, Limerick.) Cathair Chuire “close to the sea” (O’Donovan says Caher Gel Kerry, but most uncertain) ; Murbolcan (Trabolgan, Cork); Geibhtine (Askeaton, Inis Geibhthine and Eas Geibhthine, Limerick); Grafann (Knockgraffan, ‘Tip- perary); Aill micCuirr (Father Hogan says near Knockgraffan, probably only from its following that place, but no order is observed in the list); Magh Naei ; Magh nEadarba, Uacht Magh; Caechan Boirne “firm the road for the king” 1This, with Aenach mBearrain, would claim a fort at each end of the Corcavaskin territory, as the ring of forts claimed in the Corcomroes could hold that tribe in check, had the claim been reduced to reality. 2The Dinsenchas gives a Mag Findabrach (Revue Celtique, xvi, p. 69) and Brechmag, perhaps Kilfenora and Brechmag or Breffy in north-west Co. Clare, but so connected with the Meath district as to be very doubtful. For Ballykinvarga see Journal R.S.A.L., vol. xxvii, p. 121. Proc. R.I. Acad., vol. vi, ser. iii, p. 429. There are also a Kilfenora near Fenit and a Ballynavenora in Corcaguiny, Kerry ; neither has a fort of outstanding size or importance. 3 A battle was fought there a.p. 640. 4 Journal R.S. Antt. Ir., vol. xlii, p. 307. 5Tt is of unusual size among the forts of that district, measuring 235 feet over all. It has a fosse and inner and outer rings. The fort suggested by O'Donovan is not in Burrane. 6 Journal R.S.A.L., vol. xxix, p. 374 ; vol. xxviii, p. 367. 7™Mr. Goddard H. Orpen, R. 8. Antt. Ir., vol. xxxiv, p. 34, identified it with Hnach Culi in Corbally, and suggests that it is Rathmore. [3B] 20 Proceedings of the Royal Irish Academy. (Glencurraun, the “blind valley ”' of Burren, and the triple fort of Caher- commaun above the end of the “ Road of the King’s Sons,” Clare); the great Murmhagh (? Kilmurvey in Aran, O’Donovan says; some sea plain); Magh Eanaigh Rosa; Tuaim nEidhin “with its brow to the land ” (Turlough Hill fort, on the brow overlooking the O’Heidhin territory inland) ; Asal (Drum Asail or Tory Hill, Limerick); Eibhleo (Sliabh feilim, Limerick; Lissavoora Fort) ;* Ucht na rioghna (possibly Ucht na Morrighna, the Paps, Kerry); Cuilleann (Cullen, Tipperary); Cua (Sliabh Cua, Waterford); Claire (Dun Clare, Limerick) ; Inneoin (Mullaghinnone, Tipperary); Aine‘ (Knockaney, Limerick); Ord; Uilleann Eatan; Loch Ceann (perhaps in Co. Limerick near Lough Gur); Ceann Nathrach (Inchiquin Hill, Clare); the Houses of Rafann; Druim Caein; Druim Finghin “of the wood and with it Treada na riogh,” triple fort of the king (Kilfinnan ridge and the triple-fenced mote’ near it, Limerick); Rath Eire; Rath Faeladh (not Rathgel or Rathkeale as O’Donovan says): Rath Arda® (Rath arda Suird, Rathurd, Limerick); Rath Droma Deilg; Beanntraidhe (Bantry, Cork); Greagraidhe; Orbhraidhe (Orrery, Cork) ; and Ui Chuirp. In the intervening ten or eleven centuries doubtless many fort names were altered without record of the change. In later days the only important record about the Limerick forts is the repair of those of the “Islands” of Loch Gair, loch Cend, and Loch Saiglend and the forts of Dun Cliach (? Aine), Dun gCrot, Dun Eochair Maige (Bruree), and Cenn Abhrat (Claire) by King Brian 1002-1112. NUMBER AND TYPES OF THE FORTS. It is impossible to give the exact number of forts on the ground, and very hard even to decide how many are recorded on the Ordnance Survey Maps. In 1840 the latter show about 2150 in the county ; while over 1900 appear on 1So Caechan is translated locally. See Journal R. S. Anit. Iv., vol. xxvi, p. 154. Proc. R.I.A., vol. vi, ser. iii, p. 439. ? Journal R.S.A.I1., vol. xxxv, p. 224; xxxvii, p. 405. 3So identified by Rev. J. F. Lynch with great probability in North Munster Arch. Soc., vol. i, p. 114. Ebliu, daughter of Ghuaire, gave her name to the hill (Mesca Ulad, p. 149, Dind Senchas of Loch Neagh, Revue Celtique, vol. xv, p. 150). Itis ona spur of Slewphelim, and parts of its ramparts remain. The dolmen, Tuamanirvore, near it is also known as Guaire’s grave (O.S. 6, 7). * Aine and Aoife were daughters of the Sea-God, Manann4n mac Lir. 5 There are said to have been five rings, but the name Treada and the lack of any trace of the alleged two outer rings discredit the statement. ®°This fayours an early date for the list; the Norse name Siward was affixed to it in pre-Norman times. The rath is barely traceable near the curious round castle of Rathurd. The Four Masters attribute it to one of Heber’s chiefs in zB.c. 1700. Wesrropp—Larthworks and Ring- Walls in Co. Limerick. 21 the new maps. In every place known to me, many are omitted. Roughly speaking, there is a fort to about every 317 acres of land; this is regarded by some as important, but means little, as the distribution is most variable. RinG-Forts.—As to their size, the largest is the great ring, mounds and wet fosse, in which Cloncagh Church stands.!. The enclosure is 750 feet to 770 feet across, and has a well in the garth and two near it (O.S. map 37). Several forts from 450 feet to 400 feet remain—Rathanny, 450 feet, with a small central mound and two outer rings (40); Dun Clare, 450 feet to 400 feet (49) Ballykinnaugh, 420 feet by 300 feet (47); Croaghane, near Foynes, now levelled, was 470 feet across east and west, and 390 feet north and south (10). The following are about 400 feet across :—Kilmacat (4) ; Ballinscaula (40); Portauns, a D-shaped fort (47); Greenish Island (10); Killeen, near Springfield (54) ; Dromin, or Gortroe (36); Boheyglela (80). Forts 850 feet to 300 feet in diameter—Graige, oval, 350 feet to 250 feet (88); Ballinscaula, 350 feet, with small central mound (40); More- nane, shield-shaped fort (36); Cottage, of two rings (40); Tullymacthomas (45) ; Crean (31); Dunganville (28); and Glenma (39). Forts 300 feet to 350 feet across—Feohanagh, D-shaped (45) ; Lotteragh, near the castle (45); Coolrus (36); and the long oval fort of Toberyquin. Forts 250 feet across—Lisduff in Gardenfield (54); Shanid, Lower (19) ; Garranroe (21); Lisnagilla (28); Crean (31); Ballyclough, two forts (38) ; Ballylinan (44) ; Ratheahill (44) ; Kilballymartin in Mundellihy (45). Lastly, those 200 feet in diameter are—Cloneriffa (45) ; Arnacrohy (36); Clogher (89); Ballyagran (40); Mount Blakeny (47). The rest are usually about 100 feet across, or from 60 feet to 150 feet; some mere house rings, 40 feet to 50 feet across. ConsoINED Forts.—Forts of this curious type, like the Forrach and Teach Cormaic at Tara, are scarce in Co. Clare, but abound in Co. Limerick. This is merely a tentative list—Cloughkeating (13); Griston Hast (49); Incha- comb (57); Boherygella (31); Raheenamadra (41); Knockaunaskeagh or Garrankeagh (47); Baunteen (59); Ballynemore (59); Ballinscoola, two very close, but not joined (32); Clogher East, two conjoined and two very close (39); Ballykenny, two conjoined forts and an irregular fort with a D-shaped annexe (44); three are conjoined at Doonainy, on Knockaney Hill (31), Doonakemna, not in line (35), and Cush near Kilfinnane (48). The last (as we shall see in a later section) is a really remarkable group of eleven small forts between Glounacroghery Glen and Kilmurry Bridge near Moorestown. 1 Journal R.S. Antt. Ir., vol. xlii, p. 251. 22 Proceedings of the Royal Irish Academy. CRESCENT Forts.—It is hard to say with assurance that a fort was originally of this plan, but Dunganville is probably such, also, perhaps, Mountplummer (53); Ardpatrick Glen (56); Lisnafulla (53), and perhaps Dungrot. Some may be reduced to this shape by the falling away of the bank or cliff on which they abut when it is undercut by a stream. STRAIGHT-SIDED Forts.—These are not numerous, and are possibly late, if not Norman, in origin. It must, however, be remembered that they occur in Ireland far outside the English settlements, and that some in Austria date back to the Bronze Age. The chief examples in Co. Limerick are— Cowpark (small) and Shanbally (11) ; Monemoyhill (16) ; Ballycullane Upper, near Glin (17); Ballymartin (22); Cloghoonaowney in Roxborough (13); Shanid Upper (19); Garranard (two), Mundellihy, Adare (21); Skagha- phreegaun, Ardroe (23); Garrynagoora (23); Rathjordan (23); Rathbaun, Ballyhurst ; Kilduff and Gortanvally (Dromkeen South) (24); Reens (28); Duckstown and Ballywilliam, near Rathkeale (29) ; Bohora, near Anhid (30); Rathmore South (31). Lisheen, Ballynanty, consists of three small courts and a large one adjoining, 400 feet x 240 feet over all (31); Ballyrooga; Rathfreedy and Ballybrown (37); Fort Middle, Colmanstown (46); Ballin- vreena ; Garryderk; two in Glenlara; Ballygillane and Ballyfroota (48) ; Glenlara (49) ; Kells, 200 feet x 150 feet (54); Jamestown, 250 feet x 290 feet, and perhaps Ballymacshaneboy (55) ; Lackendarragh, two courts adjoining at one corner; Carheen and a second fort in Cullane (57); Bawnlogher, very small (57); Killeedy Castle (44), and Ballyfeerode, the last a ring-fort with a diamond-shaped annexe. HicH Mores.—Shanid, the castle and mote, are described below, also the second platform-fort (19) ; Kilfinnane (56), with three rings, once (it is said) five, and Grian (24), a shapely mote near Pallas Grian. None of these were called motes till recent times.‘ Small and lower earthworks are the motes of Knockaunacumsa, Dlaunaholata, and Millmount at Kilmallock, a low, small platform ; the railway runs through it. The motes of Knocklong and Bulgaden and the Eagle Mount of Bruree may also be named. The list is probably incomplete. Some are probably sepulchral—for example, Eagle Mount has no fosse, and a small pillar rises on its platform, but only excavations can distinguish between the various uses of the mounds. Crannocs.—One remains about a mile east from Rathkeale, near the railway, and several on the eastern shore of the lake, below the modern Castile of Dromore. Marsh forts are more common. The two islands of Lough Gur are recorded as island forts; one appears as repaired about 1002, but neither 1 Mote in Limerick and Clare is usually a low fort. Wesrropp—Larthworks and Ring-Walls in Co. Limerick. 28 has been explored. Lough Cenn, where another island fort was repaired, is supposed to be a marshy hollow, south from Lough Gur, but I find no evidence for this. Pintars, WELLS, AND SOUTERRAINS IN Forrs.—Ballycasey (12) and Moorestown (48) have souterrains; two were reputed to exist at Attyflin liss. Knockegan, near Askeaton, has an alignment of pillar-stones near it three in number. Bronze Age ornaments were, I am told, found not very far away. Lisnacatha is said to have had five fallen pillars in it. I have not seen it. Longstone, in Grian, has a pillar 9 feet 6 inches high in a fort. Dunmoylan, near Foynes, had two fine pillars about 10 feet high in it, but they were thrown down and broken up. If the Mesca Ulad be correct, a group of pillars stood before the great fort of Temar Luachra, and a notable pillar group on Knockaney Hill, supposed to be the echlasa, or snow shelter of the horses of the Ultonians.! Wells occur in forts at Tobermelly; Cloncagh; St. Brigid’s Well in Shangarry (36); Lisaniska (36) ; Paradise Well in Baurnagurraby fort (57). Carheen has a spring at its south-west angle. Very many others have these and wet fosses, such as the “Island” fort at Attyflin; the neighbouring Springfort and Dunganville. Clonkagh has also a stream into its fosse. A dolmen (of which only a few stones now remain) is in the works of Badgersfort, near Kilpeacon (22). Fratures.—Apart from the above and fosses and gangways, few features remain. A wall in two sections and the foundation of a gate at Ballylin are the only ones in a cathair known to me. Most of the ring-walls I have examined are absolutely featureless. DUNGANVILLE (Ordnance Survey No. 28). Turning aside from the little village of Ardagh, towards the green slopes of Sliabh Luachra, we pass below Reerasta fort, and come to a bridge and wide gully which shows with what mighty floods the Daar River ran down its gorge when no feller had come up upon the endless oak forests of Luachair, and untold acres of mist-condensing woods caught the moisture of the Gulf Stream. It was 150 feet wide, but was a narrow, shallow stream, choked with iris and ragged robin on the glorious June day of my visit in 1913. The old northern bank rises abruptly for 20 feet to the ditch of the dun, and twice as high to its platform. The name implies that a Bile, or venerated tree, grew near the fort, as at Altavilla farther westward. O'Donovan regards Emlygrennan as mbili ghroidhnin, called after some 1 Mesca Ulad, Todd Lecture Series V, vol. i (1889), R.I- Acad., vol. i, p. 17. 24 Proceedings of the Royal Irish Academy. sacred tree; but this is more than doubtful, and seems to rest (like too many of his derivations) on a guess, for in the earliest documentary form the confirmation of Prince John’s grant to Monasteranenagh, in 1199, calls it Imelach dregingi. Crecora is a more probable example; it is Craebh cumhraide according to Dr. Joyce, who, however, nearly always preferred the local to the record form. Still it is doubtful whether even the local form was con- sistent, for, in 1839, it was Craobh comhartha, and interpreted “bush of the token,” and the site of the venerated hawthorn bush, 300 feet north from the church, was pointed out. These di/eda frequently grew on or near forts® as at Magh Adhair, Roevehagh, Craebh Thelcha, and Tullaghog, and even in Christian times were objects of veneration, and centres of ceremonial, assembly, and even devotion, still maintained for trees and venerated thorn bushes in some forts in Munster and Leinster.* The fort may have been a chief “port” and inauguration place of the Ui Chonaill, as Magh Adhair was of their over-chiefs of the Dal gCais. The earliest record of the name known to me is in one of the Inquisitions, taken after the death of Thomas fitz Maurice, on July 5th, 1298, giving the manor of New Castle in Oconyl. There we find a name group Addouan, Glengort; Rouscath mor and beg; Berne 60s. ; Asdare 40s.; Donkonewall 13s. 4d.; Lystenbretenauch (Walshes- town) 40s., &c.; here the form Donkonewall is evidently a wild attempt to reproduce the unfamiliar Irish name.‘ The name appears as Donnwyll in 1452, in the Geraldine Rental of Oconyll, in OBaithin, with most of the others, notably Asdare (Asteragh in 1299), and Dowathkatyn, and again as Downgonewoolly in Peyton’s Survey of the confiscated estates of the Earl of Desmond in 1586 (p. 117b), Glandowngonwell wood being also named (p. 118), with Astaregh, Glenestary, and Doacatteen, all in Toghe Meaghan Woughtragh. Rushkeighmore and begg and Downegonewylhy were parts of Castlenoa or Newcastle granted by the Crown to Sir William Courtenay of Powderham Castle, Devonshire, in 1591.6 In the Civil Survey about 1657 (I.B. 11. 22) George Courtney held (p. 6) Doonigoniweele, and the Courtney 1 O’Hudhrin’s Topographical Poem ; see also Ord. Surv. Letters, vol. ii, p. 387. 2 In ‘‘ The King and the Hermit ”’ Tract, 44g, 101, T.C.D., No. 8 (ed. Kuno Meyer), p- 12, Marban describes the hermitage as near a ‘‘ Bili ratha,”’ venerated tree of a rath. 3 «*They cut down the Ruadh Bheiteach, and demolished its cashel,” 1143 (Chron. Scotorum, &e.). 1099, Craebh thelcha (tree of inauguration of the chiefs of Uladh at Creive, Antrim) cut down. ‘‘ The Neids were near it, in a longport” (fortress) (Ann. Loch Cé). Holy trees grow in forts of Forenaghts, Kildare, and Skeaghavanoo, Co. Clare. 4C.D.L, vol. iv, p. 257. 5 Fiants, No. 5586. Wesrropp—Larthworks and Ring-Walls in Co. Limerick. 25 Rentai (now preserved at Newcastle), in 1701 (p. 4), calls it Dongonyweell, granted to Mr. Edward Tanner, for £30 per annum, with the manor dues of wheatmeal, oatmeal, 1 hogg, | mutton, and to keep 12 men for the public service, with 12 garrons. The heriot was the best beast, or £4. Evidently Dunganville was a place of some importance on the Manor of Newcastle. The fort! is carved out of the bank of the river, and is well preserved ; it was probably crescent-shaped in plan from the first. The central fort is nearly circular, with two rings and fosses abutting on the bank, on the edge of which is left a causeway, showing that the ditch went no farther, and gives no evidence of the cutting away of the bank by the stream. The map of 1839 shows its outer ditch as circular,? but, so conventional is the marking of forts on those maps, I see no reason to regard it as true. There seem to be no existing reaches of stone facing, but other forts of Connello are nearly all faced, and the great steepness of the banks (though the earth is tough clay) may prove that such facing formerly existed here. The rings are thickly covered with hawthorns and elder, and the only entrances through the outer ring and fosse are by irregular cattle paths. The outer ring is only 4 feet high, but looked far loftier from its high bracken and plumes of foxglove. It is 10 feet thick. The fosse is from 6 to 10 feet wide below, and 4 to 5 feet deep. The next ring is 9 to 10 feet thick, and about 5 feet high over the outer fosse, but 13 feet over the great inner one, and 15 feet thick at its base. The main fosse is 15 feet wide below, and (save where filled up to the west by the inner rampart) it is almost uni- formly 13 feet deep, and is wet even in dry weather. Where it is partly filled it is 6 feet deep for a short reach, but rapidly deepens to 9 and then to 12 feet. Itis from 42 feet to 45 feet wide at the field level. The garth is 16 feet above it, the summit of the central rampart 20 feet to 26 feet, or 4 feet to 10 feet higher than the garth. It is 26 feet thick at the platform, and 9 feet on top. ‘The interior is 140 feet across inside, and 174 feet north and south, to 199 feet east and west, over the rampart. The whole earthwork measures 325 feet east and west, and 252 feet north andsouth. In the centre isa low mound of stones 9 feet wide, its south wing 27 feet, and its east wing 57 feet long, L-shaped in plan. A similar mound, parallel to the south wing, and 21 feet from the east one, lies to the north. The peasantry believe that there are caves under these, but that they were never open in the oldest memory. From the summit is the fine outlook of those gentle hills westward and to the ring of blue mountains in Clare and east Limerick — 1 Plan, Plate IV. 2 Tf correct, this would be very like Glenfoyle rath (Kilkenny Soc., R. 8. Antt. Ir., vol. 1, p. 246), which, however, had no river to cut away the lower garth. R,I,A. PROC., VOL. XXXIII., SECT, C, [4] 26 Proceedings of the Royal Irish Academy. Slieve Bernagh, Thountinna, on Lough Derg, Kimalta, and Knockfierna, raising its cairn-crowned dome and long ridges to the south-east. The fort being on the east slope of Luachair, probably once among oak groves,' and with its high rampart, deep ditches and alleged souterrain, tallies closely with the features of a famous fort in a wild romance. We may well ask was it Temhair Luchra,? the fort of Curoi mac Daire ? O’Donovan’s identification of Temar with “Bealahantouragh,” near Castle Island, does not agree with the Mesca Ulad.s There is a ring-fort in Ballahantouragh (Kerry 0. S. Map 40), about three miles east from Castle Island, but the slopes face westward, the place is shut in, and nothing in the minute topography of the Mesca tells of any long march through the hills many miles away from the sunward slopes of East Luachair. No plain exists in its valleys such as the Mesca Ulad presupposes. In brief, if it or Conreagh Fort (quasi Conroi!!), Tonreagh on the old maps, be Temar, then, unlike most of our legends, the topography of our only detailed document is radically wrong. To give one single example, the hills rise 1345 feet to the north-east. and 400 feet to the south, and 879 feet to the east of Ballahantouragh, which is about 250 feet above the sea, instead of their slopes facing the rising sun. Irish writers are hopelessly addicted to repeating outworn catchwords ad infinitum. Hennessy’s identification is far more probable, but still the Mesca Ulad does not quite tally with Portrinard, which his equation of the legendary tale with the note in the Annals of the Four Masters and the letter of Pelham compelled him to adopt. All the allusions to this Temar are extremely vague. Pelham, in 1580, marched towards Newcastle, and on to Glin, thus passing before (if not near) Dunganville. From Glin he passed through Glenlogher, and camped at Dowau. The Annalists say: “proceeding to Temar Luachra.” All turns on whether they had more than a general idea as to the position of Temar, or 1 As shown by ‘‘ Daragh”’ names near it. 2 Like Bregian Tara, it was named, according to the Dindsenchas (Revue Celtique, vol. xv, p- 444), after Tea, daughter of Lugad, son of Ith, wife to HEremon. It was the burial-place of the Kings of the Clann Deadad. 3 Mesca Ulad (ed. Hennessy). The curious lecture on Co. Limerick topography, delivered by Cuchullin, on Knockainy Hill, is on p. 17. 4 However, he states cautiously that it was ‘‘ ina triangle, the base of which ” extended from Newcastle West to Duagh, the apex being at Glin. The Onomasticon Goedelicum accepts O’Donoyan’s identification without giving anything to explain the entire contra- diction between it and the Mesca Ulad. After leaving Portrinard Pelham passed on to Tralee (letter in Carew Calendar, March, 1580, p. 236, vol. ii). The Castle Island way seems a meaningless and unnecessary circuit, and unlike other marches of the period (past Licksnaw). He more probably went past Listowel and Abbeydorney. 5 He started from Newcastle, and ravaged as he went, all along ‘‘ the foot of Slew- logher,” till he came to Shanid. Carew Cal., vol. ii, p. 206. Wesrropp—Larthworks and Ring-Walls in Co. Limerick. 27 only knew that it lay in these hills, and used it for a “district mark.” The Cath Fintragha and the Dindsenchas are vague; the Agallamh is still more vague ;! there only remains the Mesca Ulad to help us. ‘emar, according to it, lay on the slopes of East Luachair (pp. 19, 21, 23, 25), over a glen noisy with wild fowl, with forts on a ridge beyond a stream eastward.* The Ultonians came straight from Aenach Clochair, at Monasternenagh, to it; they crossed the Maigue, but are not said to have crossed the Deel’ or the Feal, or to have ascended the great hills (p. 21). They came on the same track as Queen Medbh, from the north-east, and over the side of Ir Luachra, from the east. The sun rose over the slopes of Ir Luachra “against” (opposite to) the Dun and its side (pp. 25-27). The watching Druids fell off the high rampart, one inside, the other outside. The assailants passed up the glen (p. 27). The fort had a“tulchin” or flat top. There was a subterraneous house under wooden and iron houses (p. 41). The enemy advanced to the diss, and camped on the green to the east of it. Cuchulaind leaped on to the summit of the /iss (p. 47), and on to the bridge (we. the fosse lay inside an earthwork). Alill was on the rampart of the Dun watching them. After the capture and plunder of the Dun it was never again inhabited, down to the date of the story at least. So that all we can say is that in al! the above particulars’ Dunganville fulfils the conditions ; but whether it was of the first century before Christ, or, if so, whether it was the famous fort of Curoi, has yet to be established. At least it has the broad prospect implied in its name (Portrinard has not), and it adjoins a plain’ (and not a narrow river valley like Portrinard), as Temar is said to have done. 1 Agallamh, Silva Gadelica, vol. ii, p. 176, Caeilte and Diarmuid secure the antlers of the red stag on the open lands of Luachair to the south, and the latter brings his to Temar Luachra; p. 181, Finn comes from it to Aine, andp. 238. The latter again rather supports the statements in the Mesca Ulad that the road to it led north of Knocktfierna. If the Cladh Ruadh trench, a road from Kerry Head (Journal R. Soc. Antt. Ireland, vol. xl, p. 126), led to it, and it were not at Portrinard, the western road may have crossed the hills where the railway runs from Listowel to Barna Pass. 2 Three forts lie east of Dunganville in Kilrodane, and one in Enaghgare, across the stream. The latter name implies an aenach, or assembly, such as was no infrequent appanage of important early centres. 3 Which does not allow us to attach importance to the silence of the tale about the Feal. + Perhaps the high ground near Ardagh, if Dunganville be intended. ® Save that no pillars remain before (east) of the fort, but note removal of pillars at Dunmoylan fort, Old Abbey, and Duntrileague fort. Three stand before Knockegan fort near Askeaton, and several near the promontory fort of Lissadooneen, on the Shannon, in Kerry (see Journal R. 8. Antt. Ir., vol. xl, p. 15). 6 «The plain of Teamhair Luachra” (Mac Gnimartha Finn, Ossianic Soc., vol. iv, p- 291). ‘* Luachra was a flowery plain” (Dindsenchas, loc. cit., of Temar Luachra.) [4*] 28 Proceedings of the Royal Irish Academy. Rerasta Fort, ARDAGH (0.5. 28). Between Ardagh village and Lisbane are several forts, usually small, typical rings, with fosses, and inner and outer mounds. The Cathair, which gave its name to Cahermoyle, is a massive defaced ring-wall, rarely 5 feet high, but 10 feet to 12 feet thick, with coarse facing blocks and large filling. The earthen fort of Rerasta les on higher ground than the village of Ardagh. Itis much overgrown, and, the site being in meadows at my visit, I could only examine it from a path. A deep fosse, nearly straight, runs along the north face, and is over 6 feet deep and 10 feet wide, which, with its massive mound, show it it to be a place of importance; the east and west are levelled. The place measures about 300 feet across, being somewhat oblong, and is only interesting from the find in its rampart of the beautiful chalice which has spread the obscure name of Ardagh round the antiquarian and artistic groups of the world. The chalice, as is well known, dates before A.D. 900, and is a calix ministralis;’ it is formed of gold, silver, brass, bronze, copper, lead, and enamel; with it were found a very fine brooch and a bronze chalice. Early Irish literature, in which we rarely look in vain for light on archaeological phenomena, abounds in mention of objects found in forts and mounds. ‘he great sword, 2 feet broad, dug up in Emania in 1111;? the head-piece of Smethra, brasier of Oengus of Dun Oengusa, found in Sid Cruachain; the diadem of Loeguire son of Luchta, in the Sid Findachann and the draught- board of Crimthann Nia Nair, in the rath of Usnech, may suffice to be named. The Agallamh has also much to tell of the finding of buried treasure. To select a few extracts, a mass of rings and bracelets are found in a burial mound and the excavation of a cairn, in which a shield, weapons, and a skull were found, and gold hidden in a rath’s high fence at the base of a monolith.’ The Ardagh objects may have been hidden for safety in the Norse wars, as the shrine of St. Patrick from Armagh was brought to this county for safety, by the comharb who was captured by the Gentiles at Cluain Comharba (?Colman’s Well, in this very county), and was taken in the Danish ships to Luimneach in 845, LisBANE (0.8. 19). There is a very conspicuous, though low, hill, close to the ruined church of Kilbradran, or, as many call it, “ Kilbraydon.” Certain terraces appear on its 1From which, before the cup was withdrawn from the laity, about a.p. 1000, the deacons and congregation communicated. 2(Qhron. Scotorum and other Annals. % Agallamh, Silva Gadelica, vol. ii, pp. 126, 153, 237. Westropp—Larthworks and Ring- Walls in Co. Limerick. 29 summit, which prove to be a remarkable fort,! probably the Lisbane of the townland name, and most. inadequately marked on the maps. Kilbradran was granted in 1253 to Athassell Abbey, and is named as Kylbraderan in 1291. As the fort was faced with shining grey limestone, the name Lisbane, “white fort,” was very descriptive. From its dominant position in the plain from Ardagh to the Shannon, and from the hills of Luachair to the Deel, it was probably one of the chief forts of the Ui Chonaill ; it gives interesting evidence of extensive modifications, even in early times. The original work fe DN Uy saya ee ey Weg NM tea WIN toa CN AN \\\ \\ ( \ A elise tian Ae ASAE uy, SE QF AE wi Mgy YN \\ We WY Whey : A ay ats gees. WAY \ SS Ze \y I\\ The fort of Kilbradran, County Limerick. was a strong cathair on the summit; its rampart was of earth 15 feet to 12 feet thick, stone-faced inside and out; it is still 5 feet high, with a slight fosse, 7 feet to 9 feet wide, outside it, probably rather used to supply filling for the wall than to add to its defence. The garth is slightly oval, from 91 feet to 94 feet across. Ata lower level, from about 48 feet to 70 feet outside the central fort, was a second stone-faced mound 19 feet to 21 feet thick and 9 feet to 11 feet high, also with a fosse, only 6 feet to 8 feet wide and a few 1 First noted by J. Windele. 30 Proceedings of the Royal Trish Academy. feet deep. It has been altered; an ancient road runs through it, from near the east edge of the central fort, and goes eastward in two branches down to the plain. It is banked along the fianks of the annexe, beyond which it is not traceable. Below this road was the terraced outer enclosure, which was replaced by a later and larger one farther down the slope, about 300 feet north and south and 187 feet across: its mound is 4 feet to 6 feet high and thick, with the usua! slight fosse 6 feet wide. To the north of the main rings is asimilar annexe 118 feet across and 5 feet high. The fosses were carefully Kerbed by large limestone blocks. The mounds are dotted with a few weird old thorns leaning from the west wind. There is a fine view from Shanid to Rathkeale, with the Galtees beyond, and to the north the grey terraced hills of Mullach and Glasgeivnagh, outposts of the Burren of Co. Clare. LisMAKEERY (0.8. 19). Lismakeery was another appanage of Keynsham Abbey down to the disso- lution. Its church is first named in the Papal Taxation of 1302-7. Not to follow the history, the name appears as Lismakyre, Lisvikerry, Tasmackerxy, and Lismacdyrrye in 1591, and Lissakaire, alias Tryenlassamacdirry, in 1609. The fort is a fine earthwork on a low green hill, and is a low mote or platform 16 feet higher than the fosse, and from 12 feet to 14 feet above the field, save to the south, where it is only 6 feet high. It has no outer ring, and the inner one rises hardly 2 feet above the garth to the west and south-west, most being levelled. The fosse is 6 feet to 8 feet wide below and 10 feet to 12 feet at the field level. The fort rises very steeply 16 feet, with a base of 6 feet, or 1 foot in 23 feet to 1 foot in 3 feet ; much of the stone facing remains to the south—small, beautifully fitted masonry; the north facing has been very recently thrown down, and lies in heaps in the fosse. The platform is oval, 114 feet north and south by 102 feet east and west. The late fifteenth-century church stands beside it, making a conspicuous object from Askeaton Station. The low, broken castle is in the fields below. From the summit, Shanid Castle is just visible over the ridge of Craggs, and the great keep of the Desmonds’ Castle, with the clustering houses of Askeaton, is well seen. On the opposite side of the Craggs ridge, in Ballinknockane, is a fort of curious plan in outline like a barrel, the north and south ends straight, the sides slightly curved. It is 120 feet long north and south, and from 54 feet to 90 feet across, a few feet high ; it had once a stone wall 9 feet thick, now nearly removed. Wesrropp— Earthworks and Ring- Walls in Co. Limerick. 31 Craces Group (0.8. 19). This is an interesting group of true ringwalls on a long ridge of thicket- covered crags, like a portion of the Burren of Co. Clare. It lies in Ballylin and Craggs, running to Creeves and Ballyclough. It is best reached by the steep old road across the ridge, with Ballylin and Liffane to the south, and Craggs and Ballyclough (names most descriptive of the rocky fields) to the north. “The Caher ” of Ballylin lies nearly south of the sharp bend of the road Ny x QR. Be aK ZEEE aan Woo \ Z Saige 2. \ = \ ZZ ZZ 7% BALLYCLOUGH — = — = ns \, BALLINKHOCKAR Ballyclough House, See ? Ettana “Gif Cathair = = BZ — = ° ~ IZ — ZZ z — = EN Z =e = iZ 2 Z To Askeaton ee oa a y= = i ez = ~ GZ \S Z DEELISH WG xe LIFFANE % ¥) z, i Py a ay = : - ig z CARROWOLOL s RI y Ss Ye NS = On* e i NN a A NG Z See eas Tyee 3 *° Movena House Gite. © wx 2h i Craggs Group of Forts. in that townland. It is the only fort wall with two sections so far noted by me in Co. Limerick.! The type is believed to be the “murum duplex ” of Caesar, and occurs in counties Clare, Kerry, and Galway (in some cases even with three sections), as well as in Great Britain and France, in the Alpes Maritimes. Though only 3 feet to 5 feet high, the well-marked foundation of a gateway remains to the south-east. The passage tapers from 3 feet 10 inches outside to 4 feet 6 inches inside, as at Croaghateaun and Ballyganner South, in Co. Clare, the sides elsewhere being more usually parallel, or with an offset. The piers are of large, well-laid blocks, like the facings of the walls. The two 1 Plate ILI, No 3; and plan, Plate LV. 32 Proceedings of the Royal Irish Academy. sections measure at the gateway 5 feet 6 inches and 2 feet 10 inches, the last being outside; usually the outer is the thickest, but similar inversions are at Caheridoula and the crescent two-walled fort in Carran Valley, Co. Clare. The wall has a slight batter, usually of 1 in 6. The garth and even part of the walls are so overgrown that I cannot give the dimensions. The wall is at other points 8 feet to 9 feet thick. A second but more defaced Cathair lies to the west edge of the ridge, also in Ballylin—a mere overgrown ring of small filling, 12 feet to 15 feet thick and a few feet high, with only a few facing blocks im sitw, and early house enclosures inside. In Craggs, to the north of the road, the forts are greatly levelled. (a) The first lies in the south-west corner of the townland near a conspicuous ruined cottage, and is a mere ring of filling on a low, craggy knoll. (8) Northward from the east is an earthen fort with a slight fosse, the ring barely 4 feet high and 8 feet thick, in an almost impenetrable bramble-brake and thicket. Between these forts, in a shallow, grassy depression, one finds with surprise a tombstone inscribed: “This stone was erected by Bridget Molon.. ./ in memory of her husband James Heal/y who departed this life / Mi 18 a.p. 1791, aged/ 62 years.” Ihave found such derelict monuments elsewhere ; they were possibly cut on the spot, and for monetary reasons or procrastination never taken to the grave. (c) Farther north is the foundation of a Cathair on level crag. (D) To the N.N.W., in the next field, is another ring-wall 5 feet 8 inches thick and 5 feet high, probably later than the other forts, though with fairly good facing, larger outside (as usual) than the inner. The western half is destroyed. There are no house sites visible, so it may be a bawn for cattle. The maps show yet another ring to the north-east, which I could not reach in the thickets. From this point a grassy depression runs eastward with an overgrown ridge to the south. In it is a large temporary pool covered with unusually large-leafed “silver weed” (on my visit, September, 1908) and shimmering like water. At the north-eastern edge of the townland and the main ridge lies a large cathair (£) levelled to supply material for a boundary wall. It is a ring of fairly large filling, rarely over 3 feet high, 120 feet over all, and with remains of large fine facing, evidently the chief fort of the settlement from its size and choice site on a bold knoll with a fine outlook to the Shannon and Co. Clare. The ruin is much hidden by hazel, holly, and sloe. The trace of an old road runs southward along the edge of the ridge beside the fort on the west. (F) Another fort lost in a thorny thicket I could barely locate. Three reputed forts are on the low ridge near the old road in the south-east corner of the townland. (a) The first is a nearly levelled late cattle-pen ; (H and 1) are mere foundations, The whole group, though so defaced, is a Wesrropp— Larihworks and Ring- Walls in Co. Limerick. 33 very instructive and typical settlement, and brings Limerick into comparison with the better-preserved settlements in Clare and Galway. In Ballyclough on a furze-grown bold knoll, on the edge of the slope, another fort stands about 220 feet above the sea. It is a ring-wall, 7 feet to 8 feet thick and 66 feet over all, with two faces of coarse masonry and large filling. On the north side of the garth is a hut site, and a large slab like a dolmen cover lies on the slope to the north-west. Three loops of wall adjoin the cathair, making it like a three-petalled flower in plan. The forts near the public roads along the foot of the ridge are all greatly defaced. A small, very defaced house-ring is at the cross-road to Old Abbey. A foundation of large blocks, 3 feet to 4 feet long, lies to the south of the road to Rathkeale, near Creeves Cross. The last was once a very fine ring-wall. Fort names abound in the district—Lismeenagh, Duncaha, Bally- doorlis, Dunmoylan, Lisbane, Lissatotan, Lismakeery, Lissakettle, Lismeale, Lissard, Lisnacullia, Lisnamnaroe.! FoYNES AND OLp ApBgy Groups (0.8. 10. 19). Going southward along the hill road above Corgrig from Foynes along the ridge crowned by Knockpatrick church and farther on by Shanid Castle, we find several earthworks worthy of study. In Knockpatrick fort the garth is terraced upon the slope rising 3 feet above it up hill and 10 feet down hill to the east. This type is characteristic of the Tulla-Bodyke district in East Clare. The garth is 87 feet across the ring, 12 feet to 15 thick, with traces of a fosse of equal width. Near it is a large oval earthwork called Croaghane. The south-eastern part in Sroolane has been destroyed since 1839 by an “improving tenant.” The fort is more or less levelled, and only to the north-west is the fosse traceable. It measures 390 feet north and south and 470 feet east and west, measured on the garth. ‘Two large ring-forts, one 250 feet over all, the other nearly levelled, lie farther back in the hills in Shanid Lower, and a finely situated one about 150 feet across, 528 feet above the sea in the same townland on the summit of Knockoura, a normal ring and fosse, 110 feet inside and 200 feet over all; no fort remains in Doonskerdeen. Op ApBEY.—This group is centred by the venerable and very interesting thirteenth-century Augustinian Nunnery of St. Catherine of O conyll,? with 1 Inguisitions of Exchequer, Co. Limerick, No. 10. No. 11, ann. xxi and xxvi Eliz., 1578, 1583, mentions a fort Cahergony or Catheryon, or Corgraig. I found no trace of a ring-fort at Corgrig House. * See its history by Professor John Wardell (R. S. Antt. Iv., vol. xxxiv, p- 41, and a description, p. 53). R.I.A. PROC., VOL. XXXIII., SECT. C. [5] 34 Proceedings of the Royal Irish Academy. its fish-pond, garths, gates, bridges, and pigeon-house. The subject of my paper confines me to the forts. In the large field to the west of Old Abbey lawn, are a really curious set of enclosures! connected with two ring-walls. Unfortunately a high park wall runs through the more eastern, and was built out of its material, every trace being swept away for 7 or 8 yards beside it. Subsidiary enclosures and a late house-site lie to the east of it in the lawn. ‘The cathair had a rampart 9 feet thick, with two faces of large stones usually over 2 feetlong and 18 inches deep and high. It is 105 feet across the garth and 123 feet over all, regularly curved, and with a house enclosure to the west. Strange curved enclosures of doubtful age and at, least twice modified run westward for 244 feet ; north of them was an oval hut, 18 feet by 19 feet. About 400 feet from the east fort is another foundation of a cathair of large but coarser blocks. Its wall is only 6 feet thick (another late indication in ring-walls). It is about 80 feet north and south and 68 feet east and west, and has a curved annexe, 10 feet deep to the west. Its wall is 15 feet thick, and possibly part of an older fort replaced by the lesser ring. The blocks are from 2 feet to 3 feet long and 22 inches thick and wide. The whole network recalls settlements at Hazelwood, Rossroe, and Caheridoula in Co. Clare. In the field next the pigeon-house is a stone 3 feet 6 inches high and 3 feet by 2 feet wide. It is possibly an old bounds stone. LisNABROCcK.—“ ‘I'he Badger’s Fort.” A fine well-preserved typical ring- fort of earth thickly overgrown. It measures 81 feet inside, 117 feet across the ring and 135 feet with the fosse, and is 5 feet to 6 feet higher than the field and 11 feet to 12 feet above the fosse ; it is level with the field to the west, It is 546 feet round in the fosse, and was faced with coarse stonework of large boulders and slabs. The fosse is 10 feet to 16 feet wide below, and the outer ring 4 feet to 7 feet high inside, and about 12 feet thick, being levelled with the field. ‘he revetment rose as a ring-wall above the garth, in which its large inner blocks are in situ; part has fallen into the fosse. Like so many other forts, this illustrates the valuelessness of the oft-proposed division into forts of earth and stone. In the same field is a defaced ring-wall or rather a bawn of stone-walling to the east, but of earth faced and topped with stone to the west about 5 feet high and 12 feet thick, of fairly good masonry, partly rebuilt to protect a grove of beech trees inside. LissADINWARVE.—“ ‘I'he fortified fort of the dead.” No signs of burial are visible, nor any tradition about its grim name. A low, stone-faced earth- 1 Plate IV, Wesrropp— Earthworks and Ring-Walls in Co. Limerick. 35 ring, 90 feet inside, its ring 18 feet to 20 feet thick, and only 5 feet above the fosse, which is 9 feet to 12 feet wide. The outer ring about 12 feet thick, but spread in parts to 24 feet. Between the convent and Movannan is a fort similar to the last in every respect, but with a deep fosse, 8 feet to 9 feet deep, with inner and outer rings, 12 feet thick, and completely overgrown. It is on the edge of a marshy field, possibly once a shallow lake, with rank vegetation and little minnow- haunted brooks. Dunmoyian.—The place bore its present name Dunmoylan in 1291, being held by Raymond de Valle (Wall), and was an interesting dun 90 feet across, with two lofty pillars 8 feet to 10 feet high, leaning towards each other. They were broken up, and the fort levelled, being just traceable. ‘here was a small circle of pillar-stones (now also removed) in Old Abbey next Deelish, and as I noted a line of three pillars at the fort of Knockegan. DerrtisH.— North of the Creeves road opposite the north-east corner of Old Abbey. It is also an earthwork, faced and topped with stone. Most of this has been removed, and a massive modern ring-wall built outside the old works. There is a small limekiln in the fosse. The fort is 99 feet inside; a few heaps of stone lie half hidden in bracken in the garth. The main ring is 23 feet to 27 feet thick and 9 feet high; the fosse is 9 feet wide and 5 feet deep; and the outer ring is 7 feet to 9 feet thick. The gateway faced the south, and had a gangway across the garth. The walls had two faces with small filling. SHANID CasTLE (O. S. 19). In a study of typical forts I wish to confine myself as far as possible to their description and to such part of the history of each place as may elucidate the name and the structural history and identity of the remains. This, however, compels me to study an attempt to bring what I hope to show are entirely irrelevant documents into that history. To begin with the undoubted records, they are most scanty. The name Shanid, Sean ait, means the old residence or house-site. It first appears in history as Senati, where the Ui Fidgeinti and Ui Chonaill Gabhra, in a fierce battle, routed the Norsemen with great loss, in 839.'_ In Norman times, though records of the grants of other territories and the foundation of many other castles have reached us, none, so far, have been recovered about Shanid or the Geraldine settlement which played so great a part in the history of Co. Limerick. 1 Annals, e.g. Aun. Four Masters, Ann. Ulster, &c. [5] 36 Proceedings of the Royal Irish Academy. Before 1230, Thomas fitz Maurice held it,! and with sufficient permanence to grant part to the See of Limerick to found a convent at Old Abbey. In 1282, John fitz Thomas “held a cantred in Cuny] called Shenede.”” Only in 1298 is the actual castle named to locate a smith’s house. John de Londres was its Baily in 1346. Shanid is the leading manor in Oconyll in the valuable rental of 1452; from which time, in the Inquisitions and Surveys of the Geraldine estates, in 1584-5, and the Down and Civil Survey, in 1655-7, and every record, till the barony was broken into Glenquin and Shanid, it is always located in Connello. Despite the obscurity of its history and the sreater importance of Newcastle and Askeaton from about 1280 onward, it was recognized as the chief and oldest house of the Geraldines, and their battle-cry “Shanid aboo!” is thus marked as very old.° Mr. Goddard Orpen® endeavours to get extra light on the origin of the Castle and Manor by arguing that Shanid was the castle of the district of Fontimel. He supposes that I identified Ardpatric, in the 1199 grant of Fontimel, with Ardpatrick (now Knockpatrick) in Ui Chonaill, and connects the name Fontimel, or Fontymchill, with Tinnakilla, to the north- west of Shanid. He suggests that [, and subsequently Rev. John Begley, had no reason for locating Fontimel as round Kilmallock, except our unfounded belief that Ardpatrick was the church of that name in Coshlea. As his paper on the Limerick Castles forestalled the third part of my New- castle paper on the lesser castles and the forts,’ and I am using the notes on the earthworks in this present essay, an opportunity now for the first time arises to correct these statements and explain my position. I identified Fontimel and the Ardpatric named in its earliest grant with the district round Kilmallock, but from its mention along with Askeaton regarded the Ardpatrick Castle named in the “ Ware’s Annals” or “ Dublin Annals ” as Knockpatrick. My position in the latter case cannot be main- tained, but (had it even been correct) could not possibly be held as referring to Shanid. As we see, even in the short résumé of the history of Shanid Manor, the latter is always located as in Ui Chonaill, nevér in any other 1 Senode Church was granted, circa 1230, Black Book of Limerick, p. 1087. Old Abbey lands granted to the nuns. Sir Thomas held lands there from the Bishop of Limerick. Cal. Doc. Ireland, vol. iv, p. 259. 2 Cal. Doc. Ir., vol. iii, p. 429 (1282). 3 Thid., vol. iv, p. 258. + Carew Calendar of Papers (1580), p. 236, Pelham’s letter. 5Trish war cries were abolished by Act of Parliament of Henry VII, but such Acts had little effect. § Journal R.S. Antt. Ireland, vol. xxxix, pp. 37, 38. 7 Promised, but never redeemed, Journal R.S.A.I., vol. xxxix, p. 368. The present section (though omitting the notes on the peel towers) supplies its place so far as the earthworks are concerned ; the castle notes are given to the North Munster Arch. Soc. Westrope—Larthworks and Ring- Walls in Co. Limerick. 37 eantred. It is very probably a fortification of about 1200. As to Fontimel'— the name in the best attested, because local, records is Fontymchyll or Fontimychyll. The constituent is evidently Michael, Michil, not that of Tinnakilla, Zigh na cille according to O'Donovan. ‘Tinnakill in the 133 Rental is Kyllsynkyll.2 I based my former identification on the elaborate and detailed record of the Crown cases in the Plea Rolls of 1289, and not merely on the accidental fact that it and “ Ardpatric were connected in the grant of September 6th, 1199.” In this document,‘ Thomas Fitz Maurice was granted ‘five knights fees in the tuath of Eleuri, which is in the cantred of Fontimel,” and five others in the twath of Huamerith in Thomunt, which is upon the water of Sinan.’ Ardpatric, with the residue of the cantred of Fontemel, was granted to William de Burgh. Were it even (as I believe) absolutely certain that this is Ardpatrick near Kilmallock, and not Knockpatrick in Connello, the evidence would not be decisive, as the places might be apart, like Fontemel and Huamerith; but my identification rested on another document, which leaves one in no doubt. We have the returns of the Crown cases in 1289, the first return (27) for O Carbry and Fontymchill, the last (40) for the latter cantred alone. The roll (No. 13), though faded (and in the latter part badly injured, as too often, by ill-managed attempts to revive the writing with acid), is legible. I have to thank the Deputy Keeper, Mr. M. J. McEnery, for his kind help long since, and on re-examination for this paper.’ In this, Fontymchyll’ cantred is connected with Kilmallock and 1 The chief references, besides the grant of 1199 and the Plea Rolls of 1289, are the (Irish) Rolls Close, anno li Edw. III, No. 73; Patent, anno xx Edw. III, No. 50, anno xxxii, No. 10, anno v Ric. II, No. 167. 2 Bishop Maurice de Rupefort’s Rental, Black Book of Limerick. ° Plea Rolls No. 13, anno xviii Edw. I, mem. 27, mem. 40. 4 Rotuli Chartarum (ed. T. Duffus Hardy, 1837), anno iJohn, p. 19. ® Cal. Documents, Iveland, vol. i, No. 93, No. 95. Eleuri is perhaps ‘‘Clari,’”’ Clare, or Dun Clare in Coshlea. 6 Tt is a record of much social interest. The escapes of prisoners, helped or hindered by the Irish chiefs, on the borders, the seeking of sanctuary in the churches, and the violent deaths, are curious. Besides the usual disasters common where bridges were rare, of people falling off horses (worth 5s.) and getting drowned at Gortskathe, in mid stream, and children falling down wells, there are three cases of men sitting by fires and scalded to death by boiling pots (worth 4d.), and the ‘‘murder” of a man (Hugh White) by three sows (worth 4s.). In these hopeless times we can only wish that the Plea Rolls were properly calendared for their vivid pictures of the obscure dwellers in this country. 1 may note that capitulwm may simply be a “ chapter” or heading in the record, and not a chapter or council of the ‘‘union” of parishes, as I regarded it formerly, the words ‘‘ chapter of Fontymchyll” being equivocal. ‘TI prefer this form to Fontemel, which is evidently remodelled on the name Fontemel n Dorset. So also Escloun on the Shannon was changed to Askelon! What originated the cantred name I have no means at present to discover, unless it be Kilmihil, near to and south of Kilmallock, with a well Tobervekeel (see Ord. Survey Letters, vol. i, p- 313), not far from Isnockaunacumsa mote. 38 Proceedings of the Royal Irish Academy. Kilcoana (Kilquane) churches ; officials of Emly intruded; a robber from Adare fled through it to Cork; Eccholy McEnery took other felons who fled from it, and certain robbers took refuge with Donell O’Brien, evidently of Aherloe.!. The villate of Dermeho (Darrach Mochua) was fined for harbour- ing a man who stole two horses from one Nevin O’Cahel, and it supplied a juror, who, of course, was resident in Fontymchill cantred. When, however, we turn to’the other record (eliminating all the names in O Carbry), we have even weightier evidence, for there appear fugitives to the churches of Kil- mallock (Kylmehalloc), Effyn, Duntrileague, and Dermeho;? one of the jurors (necessarily of the district) lived at Effyn; inhabitants of Kilmallock appear as residents. Among the jurors is a Martell of the family of Mortells- town, near Kilfinnane, and a Meagh of Villa Marriott (and Kilmallock). One of the Bailiffs of Fontymchyll is Robert Fot, of a family owning Fotisland, Kilmallock Another juror, Thomas Russell, is of Sawyn (the ancient battle- field of Samhain, or Knock Sawna, at Tankardstown) and the villate of Stephen Godmond (Downgodmond‘ in Particles Parish), all appear. This seems decisive that Ardpatrick in the 1199 grant was, like all the other places, in Fontymchyll, and that Fontymchyll closely corresponded to the west part of the barony of Coshlea, with Kilmallock and part of Coshmagh west and south of the Morning Star.’ This entirely disposes of theories based on the alleged “doubtful location of the cantred of Fontymkill,” and establishes the correctness of the identifications of Rev. John Begley and myself, that “it lay to the west and south-west of Kilmallock,” and to “ the south-east” of that place. In the same way, I may add that Huamerith is not the alleged unidentified place depending on “the slight indications given by O Huidhrin,” for (as the Cathreim Thoirdhealbhaigh’ shows) it lay near Sixmilebridge in Co. Clare. I write this only to clear a point of topography ! Domhnall Carrach Ua Brian was chief of the Cuanach line late in the thirteenth century. * For these places see Proc. R.I. Acad., vol. xxv, pp. 419, 423-7, and vol. xxvi, pp. 173, 189. 3 Black Book of Limerick, p. 68, circa 1234, and Fotislac in 1290, p. 67. 4 There was also a Stephenstown in Athenessie, 1655 (Civil Survey, vol. xxv, p. 11). 5 In addition Fontymkill was one of the eastern cantreds which supplied the levy of hobilers and foot soldiers against Mac Brene of Nathirlagh (Aherlo), the western half beyond the Maigue not being assessed (Pat. Roll Irish, anno xxxii Edw. III, 1358, No. 10). 5 Mr. Orpen’s paper, loc. cit., p. 87. * «They coasted along the Cratalachs . . . entered into Ui Aimrid .. . past hazel woody Baile maoil caisil towards... Cullane,” May, 1318. It extended (as we see) to the Shannon in 1199. There was another sept of Ui Aimbrit, which, perhaps, was a colony from Thomond, or sent a colony to the debatable land at Tradraighe. It dwelt in Ciarhaighe Luachra, or North Kerry. Westropp—Larthworks and Ring-Walls in Co. Limerick. 39 and to eliminate irrelevant material for the history and origin of Shanid Castle as a Geraldine appanage from a record of fact. Shanid in Elizabethan times enjoyed the reputation of being “ Desmond’s first and most ancient house of Castle Shenet,”’ as Pelham writes in 1580. Three years later the great Desmond Roll (m 11) notes “two old ruinous castles of which one is situated on the top of a very high mount, and is girded with a barbican, which, with the castle, has almost fallen.”? Presumably this happened by a natural collapse of the unstable ground on top of the mote. The remains consist of two earthworks,’ a rath of normal type, and a high mote, with a bailey, very Norman in arrangement.’ Probably these were two Irish forts, raised and modified by the early Geraldines. The massivesimplering- tower is probably of the early thirteenth century, and it is hard to fancy how, had all the mote been thrown up after 1199,° it could have been consolidated enough to bear the weight so soon. The Tower possibly rested on an older mound, which the Geraldines had capped with new earthwork. The bailey, too, is singular. Though there was plenty of room for a larger one on the fairly level summit between the rath and the mote, it runs down a steep slope, with a perverseness more characteristic of Irish fort-makers than of Norman designers ; yet the characteristics imply, I think, an undoubted Norman origin. In various earthworks elsewhere in Ireland we have absolutely certain evidence that high platform forts were gradually raised, and that high-ringed forts were filled up inside to make such an example as the rath of Shanid. This took place in Irish districts as well as in the Norman settlements, and such modifications should always be looked for and, if present, be described. The perfect preservation of the Shanid mote and rath gives no evidence, except, perhaps, at the eastern edge of the mote summit, which may imply that the raising stopped short of that segment, but may equally have been crushed down and broken by the fall of the walls and the removal of the debris. 1 Carew Cal. Papers, 1580, p. 236. Public Ree. Office, Dublin, mem. 11, ‘‘ Duobus veter. et ruinos. castell. quorum unum situatum sup. culmine montis altissim. et circuit. barbicano quod cum castello fere cecidit.” 3 See plan on Plate IV. 4 The occurrence of two forts on a hill is common in Ireland. I find an apposite case in France (Cal. of Documents France, p. 359). Aug., 1142, ‘‘ The two Castle motes of Mount Barbe, i.e., the greater and the lesser.” ‘‘Two raths that were ona tulach” are named in the Agallamh (Silva Gadelica, ed. late S. H. O’Grady, vol. ii, p. 216). 5 A mote and bretasche were made at Roscrea so late as 1245 (Cal. Doc. Ireland) ; but such structures were long made and used. Wooden castles were taken in Co. Clare, 1558 (Carew mss., Cal. I, p. 276). The palisaded mote of Ballysonan, Co. Kildare, was stormed in 1648 (Journal R.S. Antt., 1856-7, p. 111). So late as 1654 the people of Ardscull petitioned for a grant to fortify the mote there (Journal Kildare, 1896-9, vol. ii, citing General Order Book, Public Rec, Office). 40 Proceedings of the Royal Irish Academy. The Castle mote is a beautifully shaped, conical mound,! giving in its external surface no clear evidence of having been raised at various periods. Its height is 35 feet to the north, 38 feet to the east and west, and 33 feet to the south-east. It is about 69 feet across on the level summit and 411 feet round its base, its slope rising exactly 1 in 1. The fosse is 12 feet wide in the bottom ; going round from the apparent gangway and gate to the south-east, we find it deepened from 70 feet westward, further deepened at 150 feet on the south-west side to 234 feet on the west, rising up towards the east. It is a fine and well-shaped fosse 12 feet below the berm, or terrace, 9 feet to 10 feet wide; outside this was a breastwork, now about 4 feet high, whence the outer slope, some of it the untouched hillock, falls in a steep slope to the plateau. The Keep? has a massive wall 10 feet 2 inches thick and about 30 feet to 40 feet high, of strong grouted rubble, with neat outer facing, circular inside, and polygonal outside, with shallow faces. It has stepped battlements, with arrow-slits, and the late Mrs. Morgan, of Old Abbey, remembered a small turret® on the summit, to the west, long since fallen. The tower has no vaulting, or ledges, or corbels for floors; the south-west segment is standing ; some of the rock-like masses of the rest lie on the platform or rolled down the mote. One to the north-east has part of a window; part of a second window is in the tower to the north-west. The heads were turned over small planks, not over wicker centres. There is no ramp up to it, nor any sign ofa gate. The barbican wall ran round the edge, in part actually touches the keep; the lower part had formerly an exaggerated batter to hold it back from the slope 5 but this is all quarried out; above the batter it was 4 feet thick; it is 10 feet at the base, and has an extremely narrow summit and thin-stepped battlements, with slits like those of the keep. There were probably wooden platforms inside, as otherwise soldiers could scarcely have moved on the top with safety. The wall is 9 feet 6 inches high to the platform; much of it now leans outward to an alarming degree, and is badly cracked; soon all must fall down the mote, as so much has done in the past. Where the outer ring is 1 Plate III, fig. 1. Plan and sections, Plate 1V- 2T hope some specialist will face the age of the ring-tower. The indications favour its early origin when contrasted with the Castles of Askeaton, Adare, and Newcastle. 3 This appears in alittle sketch on the Hardiman map, No. 56 Trinity College Library, circa 1590. I have seen a sketch of the feature, probably late eighteenth century, but cannot recall its owner. The exaggerated view in Hall’s ‘‘ Ireland, its Scenery and Character,” vol. i, p. 374, does not show it (circa 1840, ‘‘ Green Sc.”") ; nor is it mentioned even in FitzGerald’s and MacGregor’s ‘‘ History of Limerick.” The best printed description is in that work, vol. i, pp. 363-4, Westrropp—LKarthworks and Ring-Walls in Co. Iamerick. 41 made (and not carved out of the hillock) it is 6 feet or 8 feet high; near the gate, some trace of the stonework of a pier remains. The Bailey adjoins this to the north-east. Its garth has two terraces, a steep slope falling from the fosse-ring (there 12 feet high), the ring being 30 feet to 33 feet wide; the upper terrace is 18 feet; then there are a slope 15 feet long and a wet terrace, also 18 feet wide, covered with flaggers (yellow iris). The last terrace is raised 13 feet above the fosse in the middle, but 18 feet at the south-east corner ; here is a small mound 12 feet across, probably the base of a bretesche or wooden turret. In it is a deep cut, probably made by some persons under the common obsession of treasure-seeking. The opposite corner was formerly covered by bushes, now removed, which led me to suppose that a similar mound was concealed there; but there is none. The fosse is 9 feet wide and 4 feet deep, save at the south-west turn, where, cutting through the hill edge, it is again 13 feet deep. It runs boldly up the slope to the great fosse-ring on either flank of the Bailey, getting nearer to the field level till it runs up the ring, only marked by its outer mound, which is about 6 feet wide and 3 feet to4 feet high. The Bailey itself is 99 feet long to the north-east, and 90 feet along thesouth. There seem to be old road-tracks up the hill to the east and north-east. The mote, the rath, and another ring-fort farther away to the S.S.E. are in line.! THE RaTH is a fine earthwork,’ standing on the slightly higher southern ridge of the plateau, and affords the fine view of the Castle, here reproduced. Its platform is about 125 feet over all and 110 feet inside ; slight, low ramparts surround it, and two remarkable cross-mounds? (with a pit in the centre), which I can only suppose to have been the base of some timber structure or tower. The platform is 18 feet to 20 feet high, the sides rising 1 in 11 or 2. The fosse is 8 feet to 9 feet deep and 10 feet to 12 feet wide below. The next ring is about 10 feet high and 7 feet wide on the summit, and 15 feet to 17 feet thick at the base. The outer fosse is 21 feet wide above, 12 feet below, and 5 feet to 6 feet deep, with a slight outer ring 5 feet wide on the top. There are no houses, or hut-sites, or any ancient mounds on the plateau ‘Can they be ‘“‘the high mounds” of ‘‘Shanagolden in Connello” (Sengualan Cladhaird Ua Connaill) in Cathreim Ceallachain Caisil (ed. Bugge, pp. 30 and 87), the scene of one of Cellachan’s fifteen battles with the Danes? For a good description of Shanid, see A. Curry’s account in Ordnance Survey Letters, Co. Limerick, vol. ii (ms. 14 E 8 R.I. Acad.), pp. 30-33. Gough in his additions does not name the mote. Plate III, fig. 2. Plan and section, Plate IV. * Described in Hall’s ‘‘ Ireland,” vol. i, p. 375, as ‘‘a rather deep cut ”—a curious mistake even for that most inaccurate work. R,I.A. PROC., VOL. XXXIII, SECT. C. (6] 42 Proceedings of the Royal Irish Academy. between it and the Mote.’ The usual view from the hills of Luachair, magnifi- cent in its spaciousness, lies to the east, out to the circuit of distant hills and Knockfierna; but few outstanding features are visible. Shanagolden church and village, Rathkeale, Kilbradran, the Shannon, a rich and diversified country (more English than Irish in character), lie below it. In the estuary farther away, the tall, narrow beliry of Canons Island Abbey rises within its great ring-fort, with the Clare mountains beyond. Local tradition has little to say of Shanid, save that it belonged to the Desmonds. Romantic visitors transfer to it all the picturesque story of that powerful house and its tragic ending. To say that Shanid is nearly without history brings angry contradiction from such persons; but none have yet answered the counter-challenge and given any account but a few dry, isolated facts. The origin, history, and destruction seem equally buried in silence or forgetfulness.* 1 Peyton’s Survey, p. 99, mentions a chapel—*‘ capellam nuper edificatam cujus mura tantummodo nune remanent.” Senode was granted by Thomas fitz Thomas to the see of Limerick about 1230 (Black Book of Limerick, p. 106). No ruin or tradition is traceable. 2 My thanks are due to the late Col. Morgan of Old Abbey, Mrs. Wardell, his sister, and Professor John Wardell, who in various ways helped my work in Western Co. Limerick ; and to Mr. M. J. M‘Enery, Deputy Keeper of the Records of Ireland. Proc. R. I. Acad., Vol. XXXIII., Sect. C. Plate III. Fig. 1.—Shanid Castle and Mote from the Rath. Fig. 2.—Shanid Rath from the Mote. Fig. 3.—Double-sectioned Rampart, Ballylin Uathair. Westropp. EarrHworks anpD RinG-watis ry Co. Limericr. rie wah “AT 921d 3 AY pan’ AUS w {HERES C Ami) Byyumsseeilil many M) My = Fae My Ge D K Nin 7 YY “th My, Ay yds Hon vii AJ8aVv ato Ov3LS3WOH-LY0I “MOINAWIT “OF) NI STTVM-ONIY ANV SMMLOMHIAVG —"A LOWLST AL OMOWTyT ‘Of ‘sy[Va-Sury pur syiomygaegy peord dy, _aLOW any S YFLSVO www j AWWW ° Al Symons wm) H1VWY 40 NOILOAS we au) St NW1d 303 37V9S —————— 4334 0s fo) SNOILO3S 3HL YO4s 3790S S16) or, SLOW JH1 40 NOILO3S y, GINVHS "9 "9009 “TITXXX ‘TOA “peoy ‘1H “001g TE NOTES ON IRISH MONEY WEIGHTS AND FOREIGN COIN CURRENT IN IRELAND. By M. 8. DUDLEY WESTROPP. PLATE V. Read January 10. Published Mancu 16, 1916. As at a comparatively early period coins were thin, irregular in shape, and liable to be broken and clipped, the necessity for ascertaining their true weight arose ; hence the employment of money weights. Later on, the introduction of foreign coins as legal currency made their use still more necessary. In England various proclamations relating to money weights occur from early in the thirteenth century. A proclamation of the year 1421 directed that Bartholomew Goldbeter, John Paddeslie, and John Brerner, of London, goldsmiths, and John Derlyngton, campsor and assayer of the Mint in the Tower of London, and Gilbright Vanbranburgh, engraver in the same, should be authorized to make weights for the noble, half-noble, and farthing of gold sufficient for the several cities and boroughs, and to form ten puncheons for each weight, five of them with an impression ot a crown, and the other five with a fleur-de-lis. And in the year 1422-3 John Bernes, of London, goldsmith, was appointed by the King to make the money weights for the noble, half-noble, and quarter-noble, and to stamp them according to the statute of the year 1421. Similar proclamations were issued during the sixteenth and seventeenth centuries. One of October 12th, 1587, ordered that no counterfeit pieces of current gold coin be received, or any piece lacking the just weight. And in order to enable all persons to ascertain the lawful weight, the Warden of the Mint was ordered to prepare upright balances and true weights of every piece of gold lawfully current in the realm, to be struck with an ‘E’ crowned. With regard to Ireland, references to weights for weighing the coin do R.I.A. PROC., VOL, XXXIII., SECT. C. [7] 44 Proceedings of the Royal Irish Academy. not carry us back beyond the seventeenth century, though it is very probable, from earlier proclamations and Acts of Parliament relating to coin and to foreign coin current in Ireland, that money weights were used as early as the fifteenth century. During the fifteenth and sixteenth centuries there are several Irish Statutes relating to the clipping of coin and to the currency of foreign coins in Ireland; but I have not been able to trace any definite reference to the use of money weights at this period. The following Irish Acts refer to the currency of foreign coin in Ireland, that of 1460 being apparently the earliest on that subject :— 38 Henry VI, c. 40 (1460). That, whereas in this land foreign coins had not been usually received, to the great hurt of the said land: It is ordained by authority of the said Parliament that the Rider of lawful weight be received at and of the value of four shillings ; the Ducat of full weight at and of the value of four shillings and twopence; the Lion of lawful weight at and of the value of four shillings and twopence; the Burgundy Noble at and of the value of six shillings and eightpence ; the Crown at and of the value of three shillings and fourpence ; and the Salute of lawful weight at and of the value of four shillings and twopence. And if any of the said gold coins or the English noble, half-noble, and quadrant of gold be not of the full weight, to be abated accordingly to the rate, and so to be received. 16 Edward IV, clause 43 (1476). Whereas divers gold coins of divers lands have come into this land with divers foreign merchants, which gold is not valued or set at any suitable value in this land, to the great impoverishment of the merchants and inhabitants of the same. It is ordained by authority of said Parliament that the Rider fine and good be received and passed as current in this land of Ireland of and for the value of five shillings of the money of Ireland; the Ducat fine and good of and for the value of five shillings; the Lion fine and good of.and for the value of five shillings; the Crown fine and good of and at the value of five shillings; the Crusado fine and good of and at the value of five shillings; the Burgundy Noble of and at the value of ten shillings; and the demi-Noble and quarter of the same according to the same rate; the Salute fine and good according to the rate of five shillings. And if any of the said gold coins want any part of the weight of the right standard of the same, it shall abate so much as is wanting in the payment. 28 Elizabeth, ec. vi (1586). An Act against counterfeiting or forging such kind of gold or silver of other realms as is not the proper coin of this realm, nor current in payment within this realm. In the year 1618 a proclamation was issued authorizing, in the case of England and Wales, the Master of the Mint, and in the case of Scotland, Wrsrropp—Money Weights and Foreign Coin im Ireland. 45 Charles Dickinson, sinker of the irons in the Mint at Edinburgh, to make weights for the coins then current. No mention was made of any maker of weights for Ireland, The earliest reference to the making of money weights for use in Ireland I have been able to find belongs to the year 1632. On December 20th of that year a proclamation was issued appointing Sir Thomas Aylesbury maker of money weights for England, Ireland, and Wales. Aylesbury’s patent bears date October 20th, 1632; and by it he was appointed maker of money weights for life at a yearly rent of twenty shillings. The weights were to be ready by January 26th following, and no other kinds were to be issued after that date. (Rymer’s Foedera.) Money weights occur bearing on the obverse the number of pennyweights and grains, and on the reverse the Spanish arms. One I possess bears the name of Philip IV of Spain (1621-1665). ‘These were probably used in Ireland at this period, the weights agreeing with those of the silver dollar or piece of eight and its subdivisions. (See Plate V, No. 1.) The following Acts of Parliament, proclamations, notices, etc., referring to money weights and to foreign coin current in Ireland, are set out chronologically :— July 10th, 1641. The Lords Justices and Council to Secretary Vane. In order to remedy the absence of coin in the Kingdom, we have, after consultation with foreign merchants here and with goldsmiths, thought of issuing the enclosed proclamation enhancing the value of foreign coins. We desire the King’s advice on the matter. A valuation of foreign coins, gold and silver, to pass for current in the Kingdom of Iveland at the following rates and weights :— The golden Rider or Horseman of the Netherlands, weighing 6 dwt. 12 grs. with 3 grs. allowance, : . ; o eal 4 @ Half ditto with 2 grs. allowance. The golden Rider or Horseman of Scotland, weighing 3 dwt. 6grs. with 2 grs. allowance, ; E 3 5 2h) dbl @ The half ditto with 1 gr. allowance. The golden Albertus of Brabant, weighing 3 dwt. 14 ers. with allowance of 2 grs., 11s. Od. The half ditto with allowance of 1 gr. The golden Pistolet of Spain, weighing 4 dwt. 10 grs., 15s. Od., with allowance of 6 grs. for double Pistolet and 2 grs. for half ditto. The silver Cardeseu or quarter-Crown of France, weighing 6 dwt., 1s. 8d. The half ditto accordingly. The Testoon of Portugal, weighing 6 dwt., 1s. 4d. The half ditto accordingly. None of the above shall pass in Ireland unless they weigh as above mentioned. od 46 Proceedings of the Royal Irish Academy. The true value and present value of the foregoing coins :— True Value. Present Value. 5) ss d5 Sy is ids Rider, F els AS iL Bo Horseman, . : = O10 40 Dili Albertus, : : =O) LOR a Oe l1esO) Pistolet, E i 4 9) Te O15 0 Cardescu, : : yO. 21056 ORarS Testoon, 3 : On ene 01 4 (Calendar of State Papers, Ireland.) August 16th, 1642. Ordered that a committee of this House shall repair to the Lords Justices and make known to their Lordships the great loss received by all sorts of His Majesty’s subjects in this kingdom by the making of Spanish Ryals of eight, current here for fourteen groats, whereas the same are of much less yalue in England, and in many places not current there: and, therefore, to move their Lordships to make these Ryals current only for thirteen groats, if they have the power to do so, otherwise they will think of some way of making the same known to His Majesty, to the end that the same may be current for thirteen groats only after Michaelmas next. (Irish House of Lords Journals.) November 5th, 1652. Kilkenny. Whereas there has been a custom of late years in this country of passing current clipped English money, and likewise all Spanish money called Ryals or pieces of eight, with many other sorts of foreign coin, at a far higher rate than true; ordered and declared that it may be lawful for all persons to refuse clipped English money unless tendered according to the true value by weight. That no sort of Spanish money called Ryals or pieces of eight, nor Rix, Flemish or crosse dollars, nor any other of that kind that have usually passed at the rate of five shillings, be henceforth enforced in payment for any more than at the rate of four shillings and sixpence, and the half- and quarter-pieces proportionately. Likewise, that no Philip’s money called Ducatoons, usually received for six shillings, be enforced in payment for more than five shillings and sixpence and the half Ducatoons proportionately. That no French money called Quardeques shall be enforced in payment for or above the value of four shillings and sixpence, and no other foreign coin to be enforced in payment. (Public Record Office, Dublin.) In the year 1652 the Irish Council made several representations to England with reference to the great quantities of counterfeit and clipped English money and base Peru pieces which were brought into Ireland. Wesrropp—WMoney Weights and Foreign Coin in Ireland. 47% The Peru pieces which were current for four shillings and sixpence were upon assay found to be not worth more than two shillings and fourpence. January 29th, 1660-1. A proclamation by the Lords Justices and Council. By the King’s authorization, and for removing the evils which ensue to the country by reason of the scarcity of coin in the kingdom, we declare that the following gold and silver coins now in or to be brought into the Kingdom shall be allowed and shall pass in all payments to and from His Majesty as current money, and as if they were sterling money, at the follow- ing rates :— GOLD. Dwt. ers. £ os. d. The Golden Rider, : : > 6 12 126 The half Golden Rider, . 5 3 6 Ql 8 The Spanish or French quadruple Eistelo, Wy 3 4 «0 The Spanish or French double Pistole, 8 16 112 0 The Spanish or French single Pistole, 4 8 016 0 The Spanish or French half Pistole, 2 4 0 8 0 The Double Ducat, : o eb Ih) 018 O The Single Ducat, F ; ea: 0 9 0 The Spanish Suffrain, 5 : Sa dire i 8 6 The half Spanish Suffrain, 6 > 8 18 014 8 SILVER. The Alessio) or Sevile Piece of Hight, ) eG Oe A The Rix Dollar or Cross Dollar, ) The half do. do. o & 1B 0 2 44 The quarter do. do. | 4.16 O 1 2b The half quarter do. do. 5 | 0 0 7 The Portugal Royal, c 6 2 Ad 0 0 3 8 The half-Royal, . : : GTO) 0 1 10 The quarter-Royal, : . 5 8 le 0 011 The Ducatoon, . ; : . 20 16 0 5 9 The half-Ducatoon, : ; 6 IQ 8 O 2 103 The quarter-Ducatoon, . Ne Od. 0 1 58 The old Peru Piece and French eal 5 ly © 0 4 6 The half do. do. o © 12 0 2 3 The quarter do. do. Boe sats) @ a ae The piece commonly called the Cardescu to pass as it now does. In case any of the pieces of gold or silver made current, as above, shall want the weight therein laid down, there shall be allowance given of two 48 Proceedings of the Royal Irish Academy. pence for each grain in the gold coin, and three pence for each dwt. in the silver coin. When the defects are so made up, they shall pass as if they were sterling and current in England. Unless the above coins are of the above weight, or unless such defects be made good, nobody shall be compelled to accept them in payment. Mayors, Sheriffs, Portreeves, Bailiffs, Officers of Corporations, and Justices of the Peace shall decide any difference arising according to the foregoing rule. (Calendar of State Papers, Ireland.) A proclamation, September 19th, 1662. Recites the proclamation of January 29th, 1660, and states that although the Mexico plate-pieces, commonly known by the name of Pillar pieces, be of equal fineness and greater weight than the rest of the Mexico or Civil pieces, yet many of the inhabi- tants of this kingdom do refuse to accept the same in payment of money according to their respective values, we do, therefore, hereby publish and declare that the said pieces, commonly known by the name of Pillar pieces, and the half-pieces, quarter-pieces, and half-quarter-pieces thereof were included within the intent of the late proclamation, and were thereby made current at several rates answerable and proportionable to any other of the said Mexico or Civil pieces, Rix dollars or cross dollars, mentioned in the said proclamation, according to the respective quantities thereof. The proclama- tion further states that all the several kinds of the Mexico pieces shall pass as current money in Ireland at such weights and rates, respectively, as the said Mexico or Civil pieces, Rix dollars or cross dollars, are by the late proclamation respectively to pass. Also that no persons shall be enforced to receive any of the Mexico pieces in payment unless the whole pieces shall weigh seventeen pennyweights, and the lesser pieces in proportion. Among the manuscript letters of the Earl of Essex, Lord Lieutenant of Ireland, is one to Mr. Secretary Coventry, dated February 20, 1674, enclosing a proclamation for the King’s approval, to raise the value of the Portugal Crusados, weighing fourteen pennyweights, from three shillings and eight pence, to three shillings and ten pence; and the half Portugal Crusado, weighing seven pennyweights, to one shilling and eleven pence. (Simon, Essay on Irish Coins.) A proclamation. April 9th, 1677. Whereas we are informed that divers merchants, strangers, and others have lately brought into this kingdom several pieces of Dutch coin commonly known by the name of New Lyon dollars, stamped with a lyon rampant on the one side, and a man with an escutcheon covering his lower parts, and a lyon charged in it on the other side, and coined in the years 1674, 1675, or 1676, with the motto, “ Confidens Domino non moyvetur,” and that they have dispersed and uttered the same in payments at the rate of four shillings and ninepence. And whereas we have Wesrropp—Money Weights and Foreign Coin in Ireland. 49 caused some of the said pieces to be tried and assayed by the assay master of this city, and do find that the said pieces are worse than His Majesty’s standard of England by two ounces five pennyweights in the pound weight, and that these new dollars are intrinsically worth no more than three shillings and fourpence farthing and (in proportion to the Spanish money commonly current here) worth three shillings and nine pence; and whereas none of the said pieces have been allowed to pass as current money in this kingdom, we the Lord Lieutenant and Council declare that no person or persons shall be required to take or receive any of the said pieces in any pay- ment or payments whatsoever. 1680. A proclamation by the Lord Mayor of the City of Dublin. Whereas by Act of State made and set forth here in the Kingdom of Ireland bearing the date January 29th, 1660, the piece of eight of Mexico or Sevil, the Rix dollar and the cross dollar weighing seventeen pennyweights is to pass current payment for four shillings and nine pence, the half piece weighing eight pennyweights twelve grains to pass at two shillings and four pence half- penny, and the quarter piece weighing four pennyweights six grains to pass at one shilling and two pence farthing; and whereas by like Act of State bearing date February 3°, 1662, the piece ef eight commonly called the French Lewis weighing seventeen pennyweights is likewise to pass at four shillings and nine pence the half and quarter piece rateably as in the former to pass in like manner as in the aforementioned as by the said Act of State, relation being thereunto had, doth and may more at large appear. And whereas yet, notwithstanding the plain and positive proclamation or Act of State, through diversity of weights used by many persons for the said coins in this city, the franchises and liberties thereof, great disturbance, trouble, and loss doth arise to His Majesty’s good people, and for that it is notorious that most weights used for the said coins do exceed the standard by several grains. These are therefore to give notice to all persons within this city, the franchises and liberties thereof, that do keep and use any weights for the fore- said occasions, that I have authorized and appointed Richard Lord of Copper Alley, in the city of Dublin aforesaid, goldsmith and sworn assaymaster, to make and have in readiness for all such persons as will try the same, exact weights for the several coins according to the said Acts of State; hereby requiring the said Richard Lord that he suffer none of the said weights to pass out of his hands without first bringing them and every of them to the exact standard according to the said Acts of State, and do seal and impress them with the arms of the city of Dublin and mottoes like unto those he has already 1 This proclamation has not been found. 50 Proceedings of the Royal Irish Academy. left with me,’ which shall remain in the New Hall of the said city, to compare all others his weights by; hereby requiring the said Richard Lord to demand for each set of weights he shall expose to sale but one shilling sterling and no more. And I do hereby in His Majesty’s name strictly charge and require all manner of persons whatsoever within this city, the franchise and liberties thereof, that have or shall have occasion to use or employ weights for the aforesaid coins, that they and every one of them use no weights in paying or receiving of the said moneys but what shall be tried and examined by the standard and sealed and impressed as aforesaid, as they will answer the same at their peril. And I do hereby also further authorize and require all and every of the aldermen, deputy aldermen, grand juries, and constables of this city within their wards or franchises of this city to make due inspection from time to time in the said affair; and the rather because many complaints have been made of the deceit and fraud used by weighing of money otherwise than by the said Act of State, and the weights (with the arms and above mottoes are required); and that if they or any of them shall find any or other weights in the hands of any person for receiving or paying of such or the like money within their respective wards or liberties of this city not marked and mottoed as aforesaid, to take up and secure the said weights which are to be brought before me to be tried and examined by the said standard ; as also to bring before me the person or persons using the same, to the end there may not contrary weights be used in this city, and for so doing this shall be to them and every of them a sufficient warrant. LUKE LowrHer. (Calendar of Ancient Records of Dublin. Vol V.) The National Museum and the Royal Irish Academy possess a few weights bearing the name of Richard Lord, but most of them are dated 1670. From this it is apparent that Lord was making the money weights before the Lord Mayor’s proclamation of 1680. Money weights dated 1679 made in Cork, and bearing the Cork city arms, the number of pennyweights and grains and the name of Richard Smart of Cork, goldsmith, are also in the collection of the Royal Irish Academy. I have not been able to find any proclamation with reference to these. See Plate V, No. 10. A proclamation, June 6th, 1683. Whereas many and great inconveniences have happened to His Majesty’s subjects of this kingdom by the difference of weights that have been of late made use of for weighing of such foreign coin as hath been current here by proclamation, as also by the bringing over into 1 The motto on one of 1670 is NF apDIs NEC DEMas. See Plate V, No. 3. Werstropp—Money Werghts and Foreign Coin in Ireland 51 this kingdom great quantities of Peru pieces of eight which for some time did commonly pass without any regard to their weights, for four shillings and sixpence a-piece, which was more than their real value, and of late many of them have been refused to be taken for more than three shillings or three shillings and sixpence a-piece, which is less than their real value, whereby many of His Majesty’s subjects have been at great loss; and we, being desirous to remedy a mischief that may prove so prejudicial to the trade and traffic of this kingdom by such uncertainty in the coin now current here; have thought fit by this our proclamation to declare at what rates all scrts of foreign coin, by the several proclamations now in force, are to pass amongst His Majesty’s subjects in this kingdom, which are as followeth :— GOLD. Dwt. grs 23 Gs th The Golden Rider, 6 . : o @ 1 1 2 6 The half Golden Rider, : EG 011 38 The Spanish or French quadruple Pistole, . 1 4 310 0 The Spanish or French double Pistole, 8 14 115 0 The Spanish or French single Pistole, 4 7 017 6 The Spanish or French half Pistole, 2 3% 0 8 9 The double Ducat, i 5 2 o at Ie 018 O The single Ducat, 2 6 0 9 O The Spanish Suffrance, 7 2 1 8 6 The half Spanish Suffrance, 3 13 014 3 SILVER. The Ducatoon, : : : . 20 16 0 6 0 The half-Ducatoon, 0 . a dO) 8 0 3 0 The quarter-Ducatoon, 5.64 0 1 6 The Mexico, Sevil, or Pillar piece of Hight, te) 0 oa oO Rix Dollar, Cross Dollar, or French Lewis, The halves do. do. do. > & 112 0 2 44 The quarters do. do. do. a oh OG 0 1 24 The half-quarters do. do. 5 | 0 0 The +4; part of the French Lewis, pelea? 0 0 43 The old Peru Piece of Hight, . 0 . 17 0 0 4 6 The half old Peru Piece of Hight, 6 6 1) 18 0 2 3 The quarter old Peru Piece of Hight, . . 4 6 01 14 The half-quarter old Peru Piece of Hight, i Wie, 483 0 0 632 The Portugal Royal, . 6 6 . 14 0 0 3 8 The half Portugal Royal, 6 0 ay eh HO 0 110 The quarter Portugal Royal, . : 5 «ly 0 Oll1 = @ usy R,I.A. PROC., VOL, XXXIIJ., SECT. C, 52 Proceedings of the Royal Irish Academy. And in case any of the aforesaid pieces of gold or silver shall want of the weight herein laid down on the said pieces respectively, then allowance is to be given of two pence per each grain so wanting in any piece of the kinds of gold coin, and three pence for each pennyweight so wanting in any piece of the kinds of silver coin, and so proportionately for greater or lesser wants of weight in the said pieces. And we do hereby declare that all the aforesaid sorts of foreign coin being standing weight are to pass and are to be paid and received at the rates above mentioned, and any such coins not weighing down the scales in the weighing thereof, is not to be any cause for the refusal thereof at the rates herein above mentioned. As also that every person receiving any money shall receive by what side of the scales he pleases, if he uses those of the payer, and if he uses his own, then he is to receive by which side the payer shall think proper to direct. And to the end that there may be no uncertainty in the weights of money, we do hereby further declare that we have directed all the weights that are necessary for the said silver coin to be exactly made by Henry Paris and John Cuthbeard of the city of Dublin, the stamps to be flat and the circle to be smooth and polished, that no dust may gather in, and each weight to be stamped with the number of pennyweights it bears on one side and the crown and harp on the other side, where the same may be had at reasonable rates, not exceeding twelve pence for all the weights being eight in number, viz. : for the ducatoon, half ducatoon for the whole plate and Peru pieces and half and quarter pieces thereoi, a two-penny weight, a penny weight and a halfpenny weight, which are all that will be necessary for weighing the several sorts of silver coin that do now commonly pass in this kingdom. And we have ordered a standard of all sorts of the said weights for silver coin to be left in the hands of the respective sheriffs of the several counties of this kingdom, and also in the hands of the respective mayors and other chief magistrates of the several cities and corporations of this kingdom, by whom the same are to be left in succession with the next succeeding sheriffs, mayors, and other magistrates to the end that all differences that shall happen about any weights for money may thereby be either determined or prevented. And in case any person or persons shall either pay or receive any money by any other weight than such as shall be agreeable to the weights so to be left in the said sheriffs, mayors, and other chief magistrates’ hands, they are to be proceeded against and published according to the law as keepers and users of false and unlawful weights. Wesrropp—Money Weights and Foreign Coin in Ireland. — 53 The eight weights mentioned in this proclamation and also in that of 1698, viz. — Dvyt. Grs 20 16 for the ducatoon. 10 8 » » half ducatoon. IY 0 » » Peru piece. 8 12 Pe eehaliacdos 4 6 » » quarter do. and the 2 dwt., 1 dwt., and 4 dwt. were stated to be for the silver coins, no special weights being provided for the gold. How these and the other silver coins mentioned were to be accurately weighed is not quite clear, no separate grains being included in the set. A proclamation was issued on January 16th, 1687, reciting that published in 1683, and declaring that all the foreign gold and silver coins therein mentioned should pass within this kingdom according to the weights and rates therein specified. A proclamation by the king, March 25th, 1689. Whereas we have thought fit, by the advice of our privy council, to raise the coin of this our kingdom to a higher value; we do hereby publish and declare, by the advice aforesaid, that all sorts of coin now current in this our kingdom, whether foreign or sterling, shall pass amongst all our subjects, within this our realm, and in all payments to be made either to us or from us, according to the rates following, that is to say :— GOLD. Dvt. grs. £ os. d. The Golden Rider, 3 ; : 5 lh) 5 ee 0) The half Golden Rider, : 5 o & 012 O The Spanish or French quadruple Pistole, . 17 4 316 0 The Spanish or French double Pistole, o. & 14 118 O The Spanish or French single Pistole, Be iT) 019 O The Spanish or French half Pistole, . 6 BOG ©O-9 The double Ducat, : ; ; a hal 100 The single Ducat, 6 A i a 3 26 010 0 The Spanish Suffrance, ‘ . 0 U B 111 0 The half Spanish Suffrance, . } . 38 13 015 6 The Guinea, 1 4 0 The half-Guinea, 012 0 [8*] 54 Proceedings of the Royal Irish Academy. SILVER. Dvwt. grs. ZB & th The Ducatoon, ; 5 ‘ > AO) 116 © 6 8 The half-Ducatoon, . : : 5 UO 8 038 1 The quarter-Ducatoon, : Seno men, OT The Mexico, Sevil, or Pillar Piece of Hight, ey 170 OnenO Rix Dollar, Cross Dollar, or French Lewis. The halves do. do. do. 8 12 Q 8 © The quarters do. do. do. 4 6 @ i 8 The half-quarters do. do. do. Ww 0 O 734 The 54; part of the French Lewis, . ell 2: 0 0 6 The old Peru Piece of Hight, . . ot ellepe 2 0 4 9 The half old Peru Piece of Hight, : > By 0 2 4% The quarter old Peru Piece of Hight, . Se ee) () al Bee The half-quarter old Peru Piece of Hight, Srp tne: O O F The Portugal Royal, . 5 : a Yt 0 38 10 The half Portugal Royal, ° : ve sinO @ at iil The quarter Portugal Royal, . i > 6 iy @ i @ The English Crown, : 0 5 6 The English Half-crown, 0 2 8 The English Shilling, . @ al il The English Sixpence, 0 0 64 The proclamation then states that the allowance of two pence for every grain wanting in the gold coin, and three pence for every pennyweight wanting in the silver coin, is to be given, and that the same methods be observed in the said weights as is directed by the proclamation of June 5th, 1683. A proclamation by the king, dated May 4th, 1689, states that there is in this kingdom small pieces of silver called the French three pence halfpenny or the three and a half sous, which was omitted from the last proclamation, and declares that every such piece of silver is to pass current for three pence halfpenny. A proclamation by the Lord Deputy and Council, May 29th, 1695. Whereas the coins current in this kingdom both of gold and silver have of late, by reason of the great rise of the value thereof in other parts, been carried away in so very great quantities that it is manifest unless some speedy remedy be provided, this kingdom will be soon drained of them. And whereas the raising the value of the foreign coin of the gold and silver current in this kingdom will be the most effectual means to prevent the aforesaid mischief ; we do publish and declare by this our proclamation that the several sorts of foreign coins hereafter mentioned, now current in this ~ Westropp—Money Weights and Foreign Coin in Ireland. 55 kingdom, shall from hereafter pass and be accepted, in all payments and receipts, as current money within this kingdom at and after the respective rates hereafter mentioned, viz. :— GOLD. Dyt. grs. 43 Gh The Spanish or French Pistole, ; 2 A 8 i al © The Spanish or French half Pistole, . ley Jala 010 6 SILVER. The Ducatoon, 3 : - 2006 ONGEs The half-Ducatoon, . : 2 > i) & 0 3 4 The quarter-Ducatoon, } 5 BH 4 Q i 6 The Mexico, Sevil, or Pillar Piece of eee the Rix Dollar, Cross Dollar, and all other! 17 0 @ & ¢! Dollars, and the French Lewis, The halves do. do. do. 8 12 OQ A & The quarters do. do. do. 4 6 o1 4 The old Peru Piece of Hight, . : 6 lly © 0 410 The half old Peru Piece of Hight, . 5 fs} 1 0 2 5 The quarter old Peru Piece of Hight, . we 4G O) 2) Qe The Crusado of Portugal, : : . 10 20 0 3 6 The half Crusado of Portugal, : 5 8 LO 0) ah The proclamation then goes on to state that the allowances and the weights to be used, are to be those authorized by the proclamation of June 6th, 1683. A proclamation by the Lords Justices and Council, February 21st, 1692. Whereas many and great inconveniences have happened to His Majesty’s subjects of this kingdom by the differences of weights that have been of late made use of for the weighing of such foreign coin as is current here, and that several weights for the weighing of such coin as aforesaid have been unskilfully made, sold, and uttered by John Cuthbert of the city of Dublin (who was formerly appointed one of the persons to make them), to the great prejudice of His Majesty’s good subjects; and we being desirous to remedy a mischief so prejudicial to the trade and tratftic of this kingdom, have thought fit by this our proclamation ; and we do hereby declare that we have removed and discharged the said John Cuthbert and Henry Paris from making, adjusting, or selling any money weights, and have directed all the money weights that are necessary for the said coin to be exactly made by Vincent Kidder of Dublin, goldsmith, according to the standard lately made in His Majesty’s Mint in the ‘Tower of London, each weight to be stamped with the 56 Proceedings of the Royal Irish Academy. number of pennyweights it bears, on one side, and the King’s arms on the other side, where the same are to be had at reasonable rates, not exceeding fifteen pence for all the weights, being eight in number, viz.:—for the ducatoon, half-ducatoon, for the whole plate and Peru pieces, the half and quarter thereof, a twopenny weight, a penny weight, and a halfpenny weight, which are all that will be necessary for weighing the several sorts of the said coin that do now commonly pass in this kingdom. A set of the said weights to be lodged with the clerk of the council, another with the Chief Baron of His Majesty’s Court of Exchequer, and a third with the Receiver-General of His Majesty’s revenue in this kingdom. Any persons using any other weights are to be proceeded against and punished as keepers and users of false and unlawful weights. No person except the said Vincent Kidder is to presume to make, adjust, or sell any of the money weights to be used in this kingdom for the future, upon pain of being proceeded against with the utmost severity. All the aforesaid coins being standing weight are to pass, and that any such coin not weighing down the scale in the weighing thereof is not to be any cause for refusal thereof. The Royal arms mentioned above varied somewhat on the different sets of money weights issued from time to time, as will be seen on reference to the plate. The following entries are to be found in the Irish House of Commons Journal under the specified dates :— August 17th, 1697.—Sir Francis Brewster reported from the committee for trade that they had resolved that it is the opinion of this committee, that upon a trial had before them of the several money weights made and sold by Mr. Henry Paris and Mr. John Cuthbert, there was a great neglect and mis- carriage in them by making the weights unequal, and differing one from the other, contrary to the trust reposed in them by the Government. Ordered that John Cuthbert and Henry Paris be taken into the custody of the serjeant-at-arms for their great deceit in making, selling, and uttering false money weights. Ordered that Mr. Attorney-General do prosecute the said John Cuthbert and Henry Paris for the said misdemeanour, and that the Lords Justices be acquainted that it is the desire of this House that the said John Cuthbert and Henry Paris be discharged from making any more money weights. September 15th, 1697.—Ordered that the Lords Justices be acquainted that it is the desire of this House that their Lordships will give order that Mr. John Cuthbert be prohibited from casting, selling, or uttering any more money weights. A complaint being made that John Cuthbert had since the resolutions of this House uttered and sold false and deceitful money weights, Westropp—Money Weights and Poretgn Coin in Ireland. 57 in breach and violation of the orders of this House and great fraud of His Majesty’s subjects. Ordered nemine contradicente that the said John Cuthbert be taken into the custody of the serjeant-at-arms for his said notorious violation and contempt of the orders of this House. September 16th, 1697.—Mr. Chancellor of the Exchequer reported that Their Excellencies the Lords Justices had been acquainted with the desire of this House with reference to John Cuthbert and Henry Paris, and that their Lordships were pleased to say they would give order that the same should be done accordingly. September 20th, 1697. John Cuthbert petitioned, setting forth that through the weakness and ignorance of his wife she was prevailed upon in his absence to sell weights, and begging that he may not be punished for the mistakes of another. Ordered that the said John Cuthbert be discharged from the custody of the serjeant-at-arms, paying his fees. Proclamation by the Lords Justices and Council, June 2nd, 1701. We, the Lords Justices and Council, in obedience to His Majesty’s commands to reduce the several species of foreign coin hereinafter mentioned to the rates hereinafter set forth, do publish and declare by this our proclamation that the several sorts of foreign coins hereafter mentioned now current in this kingdom, shall from and after Friday, the sixth day of the present month of June, pass and be accepted in all payments and receipts, as current money within this kingdom, at and under the respective rates hereafter mentioned, and none other, that is to say :— GoLp. Dwt. grs. £ os. d. The French or Spanish Pistole, 0 Aas 018 6 The French or Spanish half Pistole, . iia MMP ARIS 0 9 3 SILVER. The Ducatoon, ‘ : 3 . 20 16 0 6 0 The half-Ducatoon, . : : 5 10 8 0 3 0 The quarter-Ducatoon, 6 6 0 The Mexico, Sevil, or Pillar Piece of Hight, Cross Dollar and other Dollars, and the} 17 0 0 4 9 French Lewis, The halves do. do. do. 8 12 0 2 44 The quarters do. do. do. 4 6 0 1 24 The old Peru Piece of Hight, . 0 . 17 0 0 4 6 The half old Peru Piece of Eight, : > 8 12 0 2 8 The quarter old Peru Piece of Hight, . a 011i The Crusado of Portugal, : : . 10 20 0 3 8 The half-Crusado of Portugal, : . 510 Oy 58 Proceedings of the Royal Irish Academy. An allowance of two pence for each grain wanting in any piece of the kinds of the gold coin, and three halfpence for each halfpenny weight wanting in any piece of the silver coin aforesaid. A proclamation by the Lords Justices and Council was issued on August 19th, 1708, for the more effectual enforcing the several proclamations formerly issued in this kingdom, and to regulate the weight and currency of foreign coins now current therein. They, therefore, declare that the allowance of two pence for each grain in gold coins and three halfpence for each half- penny weight in silver coins is to be given; that all foreign coins of gold or silver now current by proclamations now in force in this kingdom, being standing weight, are to pass, and are to be received and paid at the rates laid down in such proclamations ; and that any such coins not weighing down the scales is not to be any cause for the refusal thereof. An Irish Act of Parliament (8 Anne, chap. vi) was issued in 1709 for the better preventing the counterfeiting the current coin of this kingdom. Among the provisoes is one that the foreign coin that is current or may be current in this kingdom is not to be counterfeited. A proclamation by the Lords Justices and Council, July 30th, 1712. We, the Lords Justices and Council, in obedience to Her Majesty’s Order, do publish and declare by this our proclamation, that the several sorts and species of foreign gold and silver coins hereinafter mentioned, shall from and after the twelfth day of August next, pass and be accepted in all receipts and payments as current money within this kingdom at the several rates herein respectively specified and none other, that is to say :— GOLD. Dvwt. grs. £ 8. Gd. The Spanish quadruple Pistole or double ) Doubloon, The Spanish or French double Pistole, Doub- | loon, and double Lewis d’or, ( The Spanish or French Pistole, 4 0 The Spanish or French half-Pistole, 2 0 The Spanish or French quarter-Pistole, ited eed 0 4 73 6 1 3 0) 1 0 Lis 314 0 pa co © w oOo H i=) o The Moidore of Portugal, The half-Moidore of Portugal, . The quarter-Moidore of Portugal, ~ for) Wesrropp— Money Weights and Foreign Coin in Ireland. 59 SILVER. Dwt. grs £58: lds The Ducatoon, F : 3 . 20 16 0 6 0 The half-Ducatoon, . p : . 10 8 0 3 0 The quarter-Ducatoon, : ohooh, x 0 1 6 The Piece of Hight of Merieo or Seville, and Mexico Pillar Dollar, and French Lewisd’or,; 17 0 0 4 9 the Rix, Cross, and other Dollars, J The halves do. do. do. 8 12 0 2 44 The quarters do. do. do. 4 6 Oo 1 2 The old Peru Piece of Hight, . 6 s iy © 0 4 6 The half old Peru Piece of Hight, ; non 2: 0 2 38 The quarter old Peru Piece of Hight, . . 4 6 Oo 1 14 The Crusado of Portugal, ‘: F . 10 20 0 3 (0 The half Crusado of Portugal. . 0 . 510 0 1 6 The usual allowances were to be given. A proclamation by the Lords Justices and Council, July 24th, 1714. We, the Lords Justices and Council, in obedience to Her Majesty’s com- mands, do publish and declare by this our proclamation that the several new species of French coins hereinafter mentioned be from henceforth current in this kingdom, and that the same shall for the future pass and be accepted in all payments and receipts as current money within this kingdom, at and under the respective rates hereafter mentioned, and none other, that is to say :— GOLD. Dwt. grs. 8B & th The French Lewis d’or of the new species, 5 & & t 2 © The half French Lewis d’or of the new species, 214 O11 O The quarter French Lewis d’or of the new species, 1 74 O 5 6 SILVER. The French silver Lewis of the new species, . 19 15 0 5 6 The half French silver Lewis of thenew species, 9183 O 2 9 The quarter French silver Lewis of the newspecies, 4 21 O 1 4% The usual allowances for any deficiency in weight to be given. A proclamation by the Lords Justices and Council, May 5th, 1718. Recites the proclamations of July 30th, 1712, and July 24th, 1714, with the lists of foreign coins to be current, and gives the usual allowance for any deficiency in weight ; also orders that Vincent Kidder and no other to make money weights which were to be for gold and silver coins as in the proclama- tions of 1712 and 1714, the weights not to exceed fifteen pence in price for R,I.A. PROC., VOL, XXXII, SECT. C. i [9] 60 Proceedings of the Royal Irish Academy. all gold and silver coins made current by the proclamation of July 30th, 1712, and fifteen pence for all gold and silver coins made current by the proclamation of July 24th, 1714. The weights to be made up in different sets. (Public Record Office, Dublin.) A proclamation by the Lord Lieutenant and Council, January 22nd, 1725. We, the Lord Lieutenant and Council, in obedience to His Majesty’s commands, do publish and declare by this our proclamation that the several new pieces of gold coin of Portugal hereinafter mentioned be from henceforth current in this kingdom, and that the same shall for the future pass and be accepted in all payments and receipts as current money within this kingdom, at and under the respective rates hereafter mentioned, that is to say :— Dwt. grs. & & ch The new gold coin of Portugal, 3 5 lf} 400 The half new gold coin of Portugal, . 5 & 2 0 0 The quarter new gold coin of Portugal, > 4h a1 i @ @ The half-quarter new gold coin of Portugal, . 2 8 010 0 The sixteenth new gold coin of Portugal, 5 i. 8 OO) The usual allowances for any deficiency in weight to be given. (Public Record Office, Dublin.) In Watson’s Dublin Almanac for 1732 a list of the following coins, with their values, is given :— 5 Gh The Guinea, i &§ @ The Pistole, 018 6 The Crown, 0 & & The Ducatoon, Oo @ @ The forty penny Piece, 0 3 4 —together with the weights and values of the quadruple pistole, moidore, and Portugal piece, and their subdivisions as set out in the proclamations of 1718 and 1725. A proclamation by the Lords Justices and Council, September 15th, 1736. Whereas several proclamations have issued from this Board for regulating and adjusting the several weights for weighing all foreign gold coin current in this kingdom, and we being well satisfied in the ability of William Archdall, of the city of Dublin, goldsmith and assay master, have thought fit to constitute, nominate, and appoint him, the said William Archdall, to make, adjust, and sell the several and respective money weights for weighing the several and respective coins made current by and according unto the said several proclamations according to the standard of weights formerly lodged Westrropp—Money Weights and Foreign Coin in Ireland. 61 with the Chief Baron of His Majesty’s Court of Exchequer, during our will and pleasure, hereby strictly commanding that no other person whatsoever do presume to make, adjust, or sell any of the said money weights to be used in this kingdom, upon pain of being proceeded against as makers and publishers of false weights. Of which all persons are required to take due notice. (“ Dublin Gazette,’ September 18th to 21st, 1736.) This proclamation was issued on account of the death of Vincent Kidder, the following notice appearing in the “Dublin Gazette” of August 17 to 21, 1736 :—“ Vincent Kidder, goldsmith and regulator of our money weights and grains, was yesterday interred in St. Werburgh’s Church.” A proclamation by the Lords Justices and Council for regulating the value of gold coins current in this kingdom, dated August 29, 1737. Whereas His Majesty has been pleased to signify his royal pleasure that a proclamation should issue for regulating the gold coins current in this kingdom, and for that purpose His Majesty’s order in Council, bearing date at his Court at Hampton Court the 21st day of July, 1737, has been transmitted to us, setting forth that whereas the Lord Lieutenant and Council of this kingdom have repre- sented to His Majesty that there is at present a great scarcity of silver coin in this kingdom, occasioned by persons being tempted to carry it out of this kingdom to make an advantage thereof, and that the greatest part of the gold coins current here is in the two larger pieces of Portugal gold, one of which passing for four pounds and the other for forty shillings, great inconveniences and difficulties daily arise in the obtaining change for the same; and there being a disproportion between the value of the said pieces and the lesser pieces of foreign gold coin, to the advantage of the larger, the same has occasioned likewise a scarcity of the lesser pieces of gold coin, by means whereof great distress has been brought upon the trade, and particularly the linen manufacture of this kingdom ; and also upon His Majesty’s forces here ; and therefore humbly prayed that the gold coin, both English and foreign, current here might be rated at the quantity of English silver they usually pass for in England, with an allowance of some small advantage to the lesser pieces. And whereas the Lords Commissioners of His Majesty’s Treasury (to whom His Majesty thought fit to refer the consideration of the said represen- tation) have reported to His Majesty in Council that they had taken the opinion of the late master-worker and the rest of the principal officers of His Majesty’s mint thereon, who proposed that a reduction should be made in the value of the gold coins current in this kingdom, at least as low as they are in Great Britain; and that the disproportion between the larger and lesser pieces should be rectified, which said proposal being agreed to by the Lords Commis- sioners of the Treasury and approved of by His Majesty in Council, His [9*] 62 Proceedings of the Royal Irish Academy. Majesty has been graciously pleased by his said order in Council to order that the following pieces of gold coin current in this kingdom do pass in payment within this kingdom at the rates hereafter respectively specified, and that a proclamation should be issued to that effect. We therefore, the Lords Justices and Council, in obedience to His Majesty’s said order, do by this our procla- mation publish and declare that the several pieces of gold coin hereinafter mentioned shall, from and after the 10th day September next, pass and be accepted in all receipts and payments as current money within this kingdom at the several rates hereinafter specified, and none other, that is to say :— Dvwt. grs. £ os. d. The Guinea at . : i 4g and all other pieces of he! same species in proportion. The Moidore, . : ; ; > Ad PY ik gd) 8 The half-Moidore, : : : 659 eli 014 8 The quarter-Moidore, . 3 ef SEALY TOS eed The quadruple Pistole or dgubile Doubloon, . 17 8 313 0 The Spanish or Bench double Pistole or ha 8 16 116 6 loon, or double Louis d’or, The Spanish or French Pistole, 4 8 018 3 The half Spanish or French Pistole, 2 ,2.4 09 2 The quarter Spanish or French Pistole, 5 pilD OF ay The French Louis d’or of the new species, 5 5 1. (2.20 The half French Louis d’or of the new species, 2142 O11 0 The quarter French Louis d’or of thenew species, 1 74 O 5 6 The piece of new gold of Portugal, ; 5 TERI Shy & The half piece of new gold of Portugal, 9) be) 1s 10 The quarter piece of new gold of Portugal, . 4144 019 6 The half-quarter piece of new gold of Portugal, 2 74 0O 910 The sixteenth piece of new gold of Portugal, de? eae tO) ecdia lL An allowance of two pence for each grain, one penny for half a grain, and one halfpenny for quarter of a grain deficient in any of the aforesaid coinsto be given. And it is declared that the weights now in use in this kingdom, and which are agreeable to the standard remaining with the Clerk of the Council, the Chief Baron of His Majesty’s Court of Exchequer and the Receiver-General of this kingdom, and no other, except as hereinafter is mentioned, shall continue to be used for weighing the several sorts of coin above mentioned, as formerly. And for the better ascertaining the weight of the said coins, we have directed one weight of half a grain and one weight of a quarter of a grain to be exactly made by William Archdall, of the city of Dublin, assaymaster, and Westropp—WVoney Weights and Foreign Coin in Ireland. 63 that the standards of such weights, where the standards of the other weights now in use are kept, there to remain to the end, that all differences that shall happen about the said weights may be either determined or pre- vented. And in case any person or persons shall either pay or receive any of the said coins by any other weights than such as shall be agreeable to the said standard, they are to be proceeded against, according to law, as keepers and users of false and unlawful weights. And we do hereby strictly charge and command that no person except the said William Archdall do presume to make, adjust, or sell any of the said money weights to be used in this kingdom for the future, upon pain of being proceeded against with the utmost severity. (“Dublin Gazette.”) In the “ Dublin Gazette ” of October 20 to 23, 1750, a list of foreign gold coins current in Ireland is given. This list agrees with that given in the proclamation of 1737, with the exception of the values of the Spanish quadruple pistole and its subdivisions, which are as follows :— Dvwt. grs. B Bo Gb The Spanish quadruple Pistole, : 5 il 8 311 4 The Spanish double Pistole, 5 116 115 8 The Spanish Pistole, . ‘ ‘ ees 017 10 The half-Pistole, 2 4 0 611 The quarter-Pistole, 1 2 0 4 54 The Barbary chiquin to pass in Dublin at the same rate as the Spanish half- pistole. A proclamation by the Lords Justices and Council, July 8, 1751. Whereas, by a proclamation bearing date July 13, 1712, the Lords Justices and Council did publish and declare that the several sorts and species of foreign gold and silver coins therein mentioned should pass and be accepted in all receipts and payments as current money within this kingdom at the several rates therein specified, and none other, and amongst others that the Spanish quadruple pistole of gold or double doubloon weighing 17dwt. 8grs. should pass at £3 14s.; the Spanish double pistole of gold or doubloon weighing Sdwt. légrs. at £1 17s.; the Spanish pistole of gold weighing 4dwt. Sers. at 18s. 6d.; the Spanish half-pistole weighing 2dwt. 4grs. at 9s. 3d. ; and the Spanish quarter-pistole weighing ldwt. 2grs. at 4s. 73d., which values were after reduced by subsequent proclamations. And whereas it is found necessary for His Majesty’s service and the good of his subjects of this kingdom, to put an immediate stop to the currency of the said Spanish quadruple pistole, or double doubloon, and the several sub-denominations, we the Lords Justices and Council do therefore, 64 Proceedings of the Royal Irish Academy. pursuant to authority from His Majesty signified to us -by His Grace the Lord Lieutenant of this kingdom, recall and revoke the said several proclamations so far forth as the same give currency to the said Spanish quadruple pistoles or doubloons of gold and the several sub-denominations thereof, and do by this our proclamation publish and declare that no collector or officer of His Majesty’s revenue or other person or persons whatsoever shall from and after the date hereof be obliged to receive in any payment or payments the said species of coins called the Spanish quadruple pistole of gold or double doubloon, the Spanish double pistole of gold or doubloon, the Spanish pistole of gold, and the halfand quarter pistole, or any of them, anything in any former proclamation or proclamations contained to the contrary therein notwithstanding. William Archdall, maker of the money weights, died September 6th, 1751, and in the “ Dublin Gazette” of September 21st to 24th the following notice appears :— “Yesterday His Grace the Lord Lieutenant and Council of Ireland were pleased to appoint Mr. Henry Archdall to be maker of the money-weights in the room of his father William Archdall, deceased.” The following advertisements appear in Dublin newspapers :-— “By authority. Weights for the several species of gold coin current in this kingdom are sold only by Henry Archdall in Darby Square, Werburgh Street, who sells the best kind of money-scales and gives the highest price for all manner of gold.” (“ Pue’s Occurrences,” September 24th to 28th, 1751.) “Henry Archdall, Darby Square, will give £3 9s. 4d. per quadruple for any quantity of gold coin over £10; £4 2s. Od. per oz. for light guineas. He sells the best kind of money-scales, and is the only person authorized to make or sell any weights for weighing the gold coin now current in this kingdom.” (“Pue’s Occurrences,” December 7th to 10th, 1751.) Henry Archdall appears to have been discharged from the position of maker of the money-weights, for in 1760 the following notice appears in the “Dublin Gazette,” of July 22nd to 26th, 1760:—“ Dublin Castle, July 25th, 1760. Their Excellencies the Lords Justices and Council have been pleased to appoint Mr. James Warren, goldsmith, to make, adjust, and sell the several and respective weights for coins made current in this kingdom, in the room of Mr. Henry Archdall.” In the “ Dublin Gazette” of September 9th to 16th, 1760, this advertise- ment is to be found :— “Whereas the Lords Justices and Privy Council have been pleased to appoint James Warren, goldsmith, of Skinner Row, to make, adjust, and sell Westropp— Money Weights and Foreign Coin in Ireland. 65 all money-weights for weighing gold coin current in this kingdom, in the room of Henry Archdall. James Warren has now a quantity of these weights ready for sale at his shop at the sign of St. Dunstan in Skinner Row, Dublin, and to prevent any person or persons counterfeiting the same, I have put I W' on one side thereof, affixed the date of the present year 1760 thereon, and marked my grains in like manner. The said weights and grains are sold nowhere else in this kingdom.” Also in “Sleater’s Public Gazetteer” of September 24th to 27th, 1768, another advertisement appears :— “James Warren, goldsmith and jeweller and maker of the money-weights for weighing all gold coin current in this kingdom, by authority of the Government, takes the liberty to inform his friends and the public that he has removed from Skinner Row to the sign of St. Dunstan, on Cork Hill, within two doors of Copper Alley, where the public may be supplied with money-weights and all sorts of the best money-scales. He also sells gold- smith and apothecary weights.” James Warren’s name appears in Dublin Directories as maker of the money-weights until 1782. The year 1760 appears to have been the last in which a dated set of Irish money-weights was issued. The other years which have come under my notice, in which sets were issued are 1670, 1680, 1683, 1697, 1698, 1709, 1714, 1718, 1737, 17388, and 1751. The weights were invariably made of brass. A proclamation by the King, June 24th, 1774, ordered that all gold coins as set out by the Commissioners of the Treasury, July 21st, 1773, were to be broken and cut if more deficient in weight than the following :— Dwt. grs. Guineas coined since December 31st, 1771, d F OTS Half-Guineas coined since December 31st, 1771, ; 216 Guineas coined during reign of George III and thetore Jan. 1st, 1772, 5 6 Half-guineas coined during reign of George III and before Jan. Ist, 1772, . 9 9 2 14 Quarter-guineas coined during the reign of Gooree Ill ana herons Jan. Ist, 1772, . : 0 0 oes ame Guineas coined before the reign of (Géoras TI, : : 5 8 8 Half-cuineas coined before the reign of George III, 0 . 218 All gold coin more deficient in weight than aforesaid shall not pass current in Great Britain. (“London Gazette,” Jnue 21 to 25, 1774.) 1 The I and the W appear on either side of the shield of arms on the reverse of the weights. 66 Proceedings of the Royal Irish Academy. An Act of Parliament (14 George ITI, chap. 92) ordered that one guinea weight and one shilling weight and also other weights, being parts and multiples of the said guinea and shilling weights, were to be made. Duplicates of the same were also to be made, and all weights were to be regulated by the duplicates, and after having been compared and found to be just and true should be marked with a stamp or mark to be approved by the Master of the Mint. Notice was given in the “London Gazette” of December 13th, 17th, and 24th, 1774, that John Whitehurst was appointed to stamp or mark all weights for weighing gold or silver coin. The mark or stamp to be an imperial crown. Money weights dating from after 1774 occur struck with various stamps, such as a coffee-pot, a lion passant, an anchor, &c., but I have not been able to find any reference to the use of these marks, Joseph Sage was appointed stamper of weights in 1788, and perhaps a change of marks took place then. From this time onward it seems that any person could make the money weights, provided that, on being found true, they were struck with the official stamp. The names of two Dublin goldsmiths appear on guinea weights— “ John Locker, 1775,” and “ William Moore, 1 Capel Street.” Moore worked in No.1 Capel Street, from 1774 to 1781. Also in the “ Limerick Chronicle” of July 13th, 1786, an advertisement appears of Charles Harrison, watch- maker, who states he makes gold scales and weights. The following is a list of the makers of money weights for use in Ireland, as far as can at: present be ascertained :— Sir Thomas Alylesbury, : 1632 | James Warren, Dublin, 1760-1782 Richard Lord, Dublin, . 1670-1683 | John Locker, Dublin, . : 1775 John Cuthbert and Henry Paris, | William Moore, Dublin, . F 1775 Dublin, . ; : 1683-1698 Samuel Gatcheil, Dublin, > cs 1800 Vincent Kidder, Dublin, 1698 1736 Richard Smart, Cork, . 5 1679 William Archdall, Dublin, 1736-1751 Charles Harrison, Limerick, . 1786 Henry Archdall, Dublin, 1751-1760 The following notices appear in Faulkner's “ Dublin Journal,” April 6 to 8, 1775 :—“ Tower money weights, under patent of the Great Seal of England, landed this day, and to be had at Craig’s in Parliament Street, on which the public may rely with the utmost safety. Scales and beams constructed on an entirely new construction.” April 27 to 29, 1775 :—“Micheal Cormick, goldsmith, sells tower stamped weights. By royal authority.” Wesrropp—Money Weights and Foreign Coin in Ireland. 67 It was stated that large quantities of coins in imitation of those of Portugal were made in Birmingham, and an order was issued to stop all such coins, dated Dublin, March 8th, 1775. (Faulkner’s “ Dublin Journal,’ March 16 to 18, 1775.) A proclamation by the Lord Lieutenant and Council of Ireland, dated March 18th, 1775, for stopping the currency of all foreign coin in this kingdom. Recites the proclamations of 1712, 1714, and 1751. “ And whereas His Majesty has been graciously pleased to signify his royal pleasure by his order in Council, bearing date at his Court at St. James, the 10th day of March instant, transmitting to us and for that purpose that a proclamation do issue to recall and revoke all proclamations heretofore published in this kingdom so far forth as they do give currency to any foreign coin whatso- ever, and to direct that no coin be accepted after a day to be named in such proclamation, in any payment whatsoever, save only His Majesty’s coin current in Great Britain. “ Now we the Lord Lieutenant and Council, in obedience to His Maj esty’s said order, do by this our proclamation recall and revoke the said several in part recited proclamations of the 30th day of July, 1712, and of the 14th day of July, 1714, and every part thereof not revoked by the said in part recited proclamation of the 8th day of July, 1751. And we do hereby publish and declare that no collector or officer of His Majesty's revenue or other person or persons whatsoever shall from and after the date hereof be obliged to take in any payment or payments any coin save only His Majesty’s coin current in Great Britain; anything in any former proclamation to the contrary thereof in any wise notwithstanding.” (“Dublin Gazette.”) Proclamation by the Lord Lieutenant and Council of Ireland dated May 8th, 1775. “ Harcourt. Now we the Lord Lieutenant and Council in pursuance of His Majesty’s order do hereby publish and declare that from and after the 24th of June next all weights to be made use of in this kngdom for weighing the gold coin current therein shall be ascertained by the duplicates of His Majesty’s weights of Great Britain lodged in the custody of His Majesty’s proper officer appointed by His Majesty for that purpose, and shall be stamped and marked with the stamp or mark provided by the said officer, and that no collector or officer of His Majesty’s revenue or other person or persons whatsoever in this kingdom shall after the 24th day of June refuse to take in payment or payments any gold coin current in this kingdom at the rates mentioned and declared by His Majesty’s proclamation bearing date the 24th day of June, 1774, to be ascertained by the said weights and no others; and that any person receiving money shall choose which side B.J.A. PROC., VOL. XXXIII, SECT. C. [10] 68 Proceedings of the Royal Irish Academy. of the scale he pleases if he uses those of the payer, and if he uses his own, then he is to receive by what side the payer think fit to direct.’ (“ Dublin Gazette.” A proclamation by the King given at the Court of St. James, dated April 12th, 1776, was issued from Dublin Castle, April 17th, 1776. The proclamation declares that ‘‘ from May 8th no guineas, half and quarter guineas more deficient in weight than the following rates, viz. -— Guineas coined before January 1st, 1772, 5 Half-guineas coined before January Ist, 1772, : 2 16 Quarter-guineas coined before January 1st, 1772, 1 be allowed to pass as current within the kingdom of Ireland except in pay- ments to be made at the receipt of our exchequer or to collectors or receivers of our revenue there, or to such person or persons appointed by His Excellency the Lord Lieutenant of Ireland; and we do require and command our Vice- Treasurer and Receiver-General and Paymaster-General of our revenues within our kingdom of Ireland from the 8th of May to the 26th of August to take and receive in payment of our revenue and taxes such of the said deficient gold coin of our realm, so as the deficiency do not exceed the following rates :— Dwt. grs. Guineas coined before January ist, 1772, : ‘ 5 6 Half-guineas coined before January 1st, 1772, : 2 14 Quarter-guineas coined before January 1st, 1772, : 1 eT “ After August 26th guineas, half-guineas and quarter-guineas as in the first table are not to pass current unless we see fit to allow fourteen days to the collector of our revenue in Ireland for the purpose of remitting and paying such of the said deficient coin. And we do require and enjoin our Vice-Treasurer and Receiver-General in Ireland during such fourteen days and no longer, to receive said deficient coin not below the weights in the second table.” (‘* Dublin Gazette,” April 16 to 18, 1776.) ; By the proclamation of March 18, 1775, the currency of foreign coin in Ireland was stopped; but in the year 1797 Spanish dollars were made current oin, and appear to have been in use in different forms until 1819. A proclamation by the Lord Lieutenant and Council of Ireland dated September 7, 1797. ‘“‘ Whereas silver Spanish dollars stamped at His Majesty’s Westropp—Money Weights and Foreign Coin in Ireland. 69 mint of Great Britain have been issued at the Bank of England and made current therein at the rate of four shillings and ninepence British per dollar ; and whereas it is expedient that such dollars so stamped should receive a like currency in this kingdom; and whereas His Majesty’s authority for the purpose has been signified, we the Lord Lieutenant and Council of Ireland do hereby publish and declare that the said Spanish dollars so stamped at His Majesty’s mint in Great Britain do pass as current money in this kingdom at the rate of 5s. 12d. each dollar, and be taken in all payments to and from His Majesty in this kingdom. “ Counterfeiters of said dollars to suffer the penalties of the Act of the 8th year of Queen Anne (Ireland), an Act to prevent the counterfeiting the current coin in this kingdom.” (“Dublin Gazette.”) A proclamation by the Lord Lieutenant and Council dated October 19th, 1798. ‘Ordered that the Spanish dollars as current by the proclamation of September 7th, 1797, be called in and not pass as current money in Ireland. All dollars to be brought to the Bank of Ireland at the rate of 5s. 13d. during twenty-one days from the date of the proclamation, and such dollars as are not brought in said time, will after expiration thereof pass current and be received in payment at the rate of 4s. 10d. per each dollar.” This first issue of Spanish dollars of Charles II] and IV was stamped with the bust of George III. The stamp was oval, similar to that struck on sterling silver, and was stamped on the neck of the bust of the Spanish King on the coins. In 1804, as it was found that these dollars were largely counterfeited, the counterstamp was enlarged, and the bust of George III was placed in an octagonal stamp, An Act of Parliament (44 George III, chap. 71) was passed to prevent the counterfeiting of the Bank of England dollar tokens. These, together with the counterstruck dollars, were largely counterfeited in various ways. Some were forged by taking two genuine dollars, filing them down to about the thickness of brown paper, then soldering the obverse and reverse to a copper disc and plating the edge; others were made of a disc of Sheffield plate stamped out in a disc, while others again were base metal plated and stamped. The counterfeiting appears to have been done chiefly in Birmingham. See “ Numismatic Circular” (Spink), September-October, 1915. As the Spanish dollars were still largely counterfeited, they were restruck in 1804, and were slightly larger than the original dollars. Dies were prepared by Mathew Bolton of Birmingham, the obverse with the head of 70 Proceedings of the Royal Irish Academy. George III, and the reverse with “five shillings dollar” in an oval enclosing Britannia seated. The following notice appears in Faulkner’s “Dublin Journal” of July 17th, 1804:—< Bank of Ireland. Notice is hereby given that the dollars stamped into silver tokens at Mr. Boulton’s manufactory which the Bank of Ireland is now issuing for six shillings each will be received in payment again at the Bank at the same rate, provided they shall not be defaced or mutilated or any way rendered lighter except from the operation of common wear. “« By order, “Tuomas WILLIAMS, Secretary. “N.B.—The Bank reserves the power to call them in at any time upon giving three months’ notice.” These dollar tokens had on the obverse the bust of George III similar to that on the English ones, and on the reverse Hibernia seated and “Bank of Treland Token, six shillings, 1804.” Dollar tokens appear to have been in use until April 5th, 1819 (58 George III, chap. 14), the first five-shilling pieces of George III being issued n 1818. With the withdrawal of the Spanish dollars the currency of foreign coin in Ireland ceased. A proclamation was issued on July 1st, 1817, for regulating the weights for the gold coin. Those more deficient in weight than the following were not to pass as current :— Dwt. grs. Dvt. grs. Guineas, : : 5 8 Seven-shilling pieces, 1 18 Half-guineas, . : 2 16 Sovereigns, 5 23 Quarter-guineas, . 1S The half-sovereign, made current by proclamation, October 10th, 1817, was to weigh 2 dwt. 13} grs. Money weights for guineas, half-guineas, sovereigns, and half-sovereigns continued to be used, some being made in Dublin during the first half of the nineteenth century by Samuel Gatchell. The latest weights that have come under my notice are sovereign and half-sovereign weights of the Royal mint of 1843. [EXPLANATION OF PLATE. bo (SS) 10. iil, 14. Wesrropp—Money Weights and Foreign Coin in Ireland. 71 EXPLANATION OF PLATE V. Obverse and Reverse of each Weight, slightly reduced. . 17 dwt., c. 1660, for the Mexico or Seville Piece of Hight, the Rix and Cross Dollars. . Sdwt. 12 grs., 1670-80, for the Half Peru Piece of Hight. By Richard Lord. . 4 dwt. 6 grs., 1670, for the Quarter Peru Piece of Hight. By Richard Lord. . 19 dwt. 14 grs. 8m., 1714, for the French Silver Louis. By Vincent Kidder. . 2dwt., 1683. One of the three extra weights issued. . 17 dwt., 1697, for the Peru Piece of Hight. By John Cuthbert and Henry Paris. . 10 dwt. 8 grs., 1698, for the Half Ducatoon. By Vincent Kidder. . 6 dwt. 22 grs., 1709, for the Moidore of Portugal. By Vincent Kidder. . 8dwt. 12 grs., 1683, for the Half Peru Piece of Hight. By John Cuthbert and Henry Paris. 17 dwt., 1679, Cork. For the Mexico or Seville Piece of Hight, the Rix and Cross Dollars. By Richard Smart. 18 dwt. 103 grs., 1737, for the piece of New Gold of Portugal. By William Archdall. . 17 dwt. 8 grs., 1718, for the Spanish Quadruple Pistole or Double Doubloon. By Vincent Kidder. 3. 1dwt., 1697 or 1698. One of the three extra weights issued. By Vincent Kidder. 17 dwt. 8 grs., 1738, for the Spanish Quadruple Pistole or Double Doubloon. By William Archdall. R.I.A. PROC., VOL. XXXIII., SECT. C. [11] Proceedings of the Royal Irish Academy. 5. 9 dwt. 54 grs., 1751, for the Half piece of the New Gold of Portugal. By Henry Archdall. 2 dwt. 14h ers., 1760, for the Half French Louis d’Or. By James Warren. . 5dwt. 3 grs.,1751. Henry Archdall’s initials. For guineas before the reign of George III. . 5dwt. Sgrs., 1760. James Warren’s initials. For guineas before the reign of George III. 5 dwt., 1775. By John Locker, Dublin. 5 dwt. 8 grs., c. 1800. By Samuel Gatchell, Dublin. For the guinea. Proc. R. I. ACAD., VoL. XXXIII., Srcr. C. PLATE V. WESTROPP.—IRISH MONEY WEIGHTS, ETC. [ 73 ] IV. LIST OF BOOKS AND TRACTS PRINTED IN BELFAST IN THE SEVENTEENTH CENTURY. By E. R. M‘CLINTOCK DIX. [Read January 24. Published Marcu 18, 1916.] Havine dealt with the seventeenth-century printing of the cities of Cork, Kilkenny, and Waterford, I propose now to deal with that of the city of Belfast, the only other provincial town in Ireland of which there survive specimens of its printing press in that century. There is evidence of printing in Limerick and Drogheda in the seventeenth century, but no specimen is extant or at present identified. There may also have been a press printing in Londonderry for a brief period, but this is uncertain. Printing in Belfast was very well recorded by the late Mr. John Anderson, Hon. Secretary to the Linen Hall Library, who expended years of research and much money in dealing splendidly with the subject, about which he was an enthusiast, and he was the first Irish bibliographer who published an exclusively bibliographical work, 1.e., “A Catalogue of Early Belfast Printed Books,” 1890, and two supplements. From him I drew my own inspiration, and took my first model; but in his well-known work full collations are not given of any work save of one edition of the Bible. Also, since his lamented decease, some years ago, additional items oi the earliest Belfast printing have been traced, and can now be fully collated. This, then, is all that 1 propose to do in this list, but it is desirable, I think, to have the earliest Belfast printing properly collated, and the places where items are to be found again denoted. The total items in this list are eighteen in number. Details of the finding of the leaves of the New Testament (No. 17 in the following list) appear in the “Irish Book Lover,” vol. vi, pp. 159-60, and Dr. J. S. Crone, the editor and discoverer, deserves great credit for his R.I.A. PROG., VOL. XXXIII., SECT. C. [12] 74 Proceedings of the Royal Irish Academy. discernment and bibliographical skill and knowledge in connexion with this very interesting find. Through his courtesy | am enabled to reproduce a facsimile of one of the fragments so discovered by him. Mr. R. M. Young is another Belfast bibliographer who has Eonininted materially to our knowledge of the first Belfast presses. Iam indebted to him for permission to reproduce the facsimile of a title-page of a book in his possession. The splendid collection of books made by the late Mr. Lavens M. Ewart, and now resting for a while in the Linen Hall Library, Belfast, contains very rare and early items of Belfast printing. It would be a decided advantage if its catalogue were published. In the Linen Hall Library itself are many items of Belfast printing, and I am indebted to its Librarian, Mr. F. J. P. Burgoyne, for much aid in completing this list. It may be mentioned that Archdeacon Cotton gives 1696 as the date of Neill’s printing press being set up in Belfast; but this is, I think, a printer’s error for 1694, as his further statement, at p. 19, of his Typographical Gazetteer, 2nd series, shows. Mr. Anderson has also very correctly pointed out in the preface to his first supplement, that William III had an ambulatory press with his army in Ireland, and it may have been used to print proclamations in Belfast at that earlier date. All the items in this list are of a religious character, and appear to be chiefly reprints of Puritan theological works. Several have advertisements of “ books printed and sold” by Neill, and copies of these are given in Mr. J. Anderson’s Catalogue and supplements. The library of the late Rev. T. W. Carson was sold after his decease about twenty years ago, and it is not known what became of the rare items of early Belfast printing which (as shown in this list) he at one time owned. 1. 1694. The Scottish Catechism, with the Solemn League and Covenant. Stated by Archbishop King to have been printed in Belfast in this year and am 1700. [Vide Cotton’s Typographical Gazetteer, 1866, 2nd Series, p. 19. John Anderson’s Catalogue of Early Belfast Printed Books, &c. R. M. Young in ‘‘ The Library,” vol. vii (1895), p. 135.] 2. 1697. An Answer to the Bishop of Derry’s [William King] Second Admonition to the Dissenting Inhabitants in his Diocess. Especially as to Dix—Printing in Belfast in the Seventeenth Century. 75 Matters of Fact, Relating to the publick Worship of God Wherein his Misrepresentations are again Discovered. Robert Craghead. 4to. 7} x 53. 7 leaves + 1-166 pp. [Linen Hall Library, Belfast: Magee College, Derry, 3. H. 12.] N.B.—No place or printer given. 3. 1697. Animadversions on The Defence of the Answer To a Paper, Intituled “ The Case of the Dissenting Protestants of Ireland In Reference to a Bill of Indulgence, from the Exceptions made against it.” Together With An Answer to A Peaceable and Friendly Address To the Non-Conformists Written upon their desiring an Act of Toleration without the Sacramental Test. [John McBride]. 4to. 118 pp. 73x 53. Sigs. A-Z. Aa-Ff. [Assembly College, Belfast. (Imperfect. pp. 17-91.) National Library (Joly Collection—Books), cut down. Advocates’ Library, Edinburgh—perfect copy. Magee College, Derry. (Imperfect—lacks all after p. 92, blank ; much cut down.)]| N.B.—No place or printer.] 4. 1697. An Answer to a Peaceable and Friendly Address, &c. John McBride. 4to. [Assembly College, Belfast. (Imperfect. pp. 95-118). ] N.B.—-See No. 3. Query: Is above not part of it? 5. 1698. A / Sermon / before the / Provincial Synod / at / Antrim. / Preached June 1, 1698. / By / John MacBride, / Minister of Belfast, / Published at the Desire of some Persons/ then Present / Printed in the Year MDCXCVIIL./ 4to. Title leaf + 20 pp. 7% x53. Sigs. A-E. [Brit. Mus. / 4476. d. 84. Assembly College, Belfast. (Imperfect, lacks title leaf.) Marsh’s Library (Cashel Collection). Shelf VII. Vol. xix. Advocates’ Library, Edinburgh. University Library, Cambridge. / Hib. 5. 698. 3.] N.B.—No place or printer given. 6. 1699. The Psalms of David in Meeter. Newly Translated, and diligently Compared with the Original Text, and former Translations: More plain, smooth, and agreeable to the Text, than heretofore. Allowed by the Authority of the General Assembly of the Kirk of Scotland, and appointed to be sung in Congregations and families. (Patrick Neill & Company). 12mo. 5 x 2%. pp. 1-130 +1 leaf. (Advertisements. ) [In the ‘‘ Lavens M. Ewart” Collection in the Linen Hall Library, Belfast. ] [12*] 76 Proceedings of the Royul Irish Academy. 7. 1699. The Christians Great Interest, or, a Short Treatise, Divided into two Parts: The First whereof containeth, The Trial of a Saving Interest in Christ. The Second pointeth forth plainly The Way [How] to attain it: Wherein somewhat is likewise spoken to the Manner of express Covenanting with God. William Guthrie. (Patrick Neill and Company). 12mo. 43 x 23 (much cut down). 192 pp. [In the ‘‘ Lavens M. Ewart” Collection in the Linen Hall Library, Belfast. ] 8. 1699. [The Bible the Best New Year's Gift] (in Verse) “Dedicated to King William.” (Patrick Neill & Company). 37}; x 13. 34 leaves (incomplete), including woodcuts. [The John Rylands Library, Manchester. ] Nore.—The Dedication to the New Testament is signed ‘‘J. Taylor.” O.T. B83, &c. Imperfect. N.T. Sig. D3 to E6 (Acts). Imperfect. 9 woodcuts inserted. The text consists of a rhyming setting of some of the incidents, &c., of the different books. See Ulster Journal, vol. xii (1900), p. 41. The title of the New Testament runs: THE / New / Testament. / Dedicated to / King Witttiam. / Belfast, / Printed by Patrick Neill and / Company and sold at his / shop. 1699./ 9. 1700. The Psalms of David in Meeter. Newly Translated and diligently Compared with the Original Text, and former Translations. More plain, smooth, and agreeable to the T’ext, than any heretofore. Allowed by the Authority of the General Assembly of the Kirk of Scotland, and appointed to be sung in Congregations and Families. [Sir Francis Rous.] (Patrick Neill and Company.) 32mo. 45x 24. Title leaf (verso blank) + pp. 3 (Sig. A2) to 287 (verso = advertisement). [First Presbyterian Church, Belfast. ] N.B.—Bound in Tortoise Shell with Silver Mounts. 10. 1700. The Almost Christian / discovered :/ or, / the False Professor / Tried and Cast. / Being the substance of seven /Sermons,/ First preached at Sepulchres, London, / 1661, and now at the Importunity of / Friends made publick. / By Matthew Mead./ (Patrick Neill & Company.) 12mo. Title leaf + (iii)—(xvi) + pp. 17-224. 47 x 3 (cut down). [The late Rey. T. W. Carson, Dublin. The ‘“‘ Lavens M. Ewart” Collection in Linen Hall Library, Belfast. Brit. Mus. / 4474.a. 89.] Dix—Printiny in Belfast in the Seventeenth Century. 77 11. 1700. The Life, Death, and Burial of John Flavell, and two Sermons, The Character Of a Compleat Evangelical Pastor: Drawn by Christ, Mat. 24, 25, &c., and a Coronation Sermon. John Galpine. (Patrick Neill & Company.) 12mo. 3x4, [First title leaf, wanting, pp. iii-xviii (Mr. John Flavell. The Epistle to the Reader, giving a brief account of this Excellent ..... Author, h(is) Character, Life, Death, and IH os ooo HOOGs a couple of learned ... Sermons, &c., &c.) + pp. 19-56. (A Corona- tion Sermon preached at Dartmouth, &c.)—last p. blank + Title leaf (The Character of a Compleat Evangelical Pastor, &c.)—Verso blank + pp. 59-105}. [R. M. Young, Belfast. ] THE CHARACTER Of a Compleat Evangelical Paftor ; Drawn by CHRIST, MAT. 24. 45) 46, 47: Opened and applied in a SERMON intended to be preached at Tzumton, in the County of Somerfer, at the Defire, and by the Appointment of feveral United Brethren, Of Gloucefter, Dorfet, Somerfer, and De- vonfbire, at their Meeting there, Sep- tember, 1691. By JOHN FLAVELL, late Preacher of the Gofpel at Dartmouth in Devon. 3 IB JE 12 AIS 6 Printed by Patrick Neilland Company, and fold at his Shop. 1700. Fic. 1.—Facsimile of 2nd Title-page of No. 11. 78 Proceedings of the Royal Trish Academy. 12. 1700. Sighs from Hell; /or the / Groans of a damned Soul./ Discovering from the 16th of Luke, the / Lamentable State of the Damned. / And may fitly serve / As a Warning-word to Sinners, both Old/and Young, by Faith in Jesus Christ, / to avoid the same Place of Torment. / With a Discovery of the Usefulness of the Scriptures, / as our safe-Conduct for avoiding the Torments of / Hell. John Bunyan. (Patrick Neill & Co.) 12mo. Title leaf, + ii—vui, pp. 8-192. 48 x 3. [Brit. Mus., C. 58. aa. 4.] 13. 1700. The Great Concern: or a Serious Warning To a Timely and Thorough Preparation for Death; With helps and Directions in Order thereunto. Edward Pearce. (Patrick Neill & Company). 12mo. 5 x 3} (cut down). [The late Rey. T. W. Carson, Dublin ; The ‘‘ Lavens M. Ewart ” Collection in Linen Hall Library, Belfast.] 14, 1700. Time and The End of Time, in two Discourses: The First about Redemption of Time; The second about Consideration of our latter End. John Fox. (Patrick Neill and Company). 12mo. 5 x 34. 238 pp. (advertise- ment on last page). {Linen Hall Library, Belfast. ] 15. 1700. War with the Devil: or, The Young Man’s Conflict, with the Powers of Darkness ; in a Dialogue, Discussing the Corruption and Vanity of Youth, the horrible Nature of Sin and deplorable Condition of fallen Man. Also a Definition, Power, and Rule of Conscience, and the Nature of true Conversion to which is added, An Appendix, containing a Dialogue between an old Apostate, and a young Professor, Worthy the Perusal of All, but chiefly intended for the Instruction of the Younger Sort. B(enjamin) Keach. (Patrick Neill & Company.) 12mo. 5x3. 180 pp. +1 leaf (advertise- ment). [The ‘‘ Lavens M. Ewart” Collection in the Linen Hall Library, Belfast. ] 16. 1700. A Most / Familiar Explanation /of the / Assemblies / Shorter Catechism /. Wherein their Larger Answers are bro/ ken into lesser Parcels, thereby to let / in the Light by degrees into the Minds/ of the Learners. / To which is added, in the close, a most brief / Help for the necessary, but much neglected Du/ ty of Self-examination, to be daily perused./ And to 79 ‘y). the Seventeenth Centur in g in Belfast eo Dix— Printin By this is subjoined a Letter of Christian Counsel to a/ destitute Flock. Jos. Allaine. Title leaf + (iii)-(vi) + pp. 7-144. (Patrick Neill & Company.) 8vo. 48 x 92, (Brit. Mus. /3505. aa. 103: cut down.] [17002] The New Testament. (St. Mark.) Fragments. (Weill.) 8vo. [Linen Hall Library, Belfast. (Presented by Dr. J. S. Crone.)] cols to a 2 17. page. 6 x 4. ‘prey 24 pay or {Hoom 1(0 Jur find ‘sy3aesp ay) o1ut 0 yI208 pre ‘h179q oy) Ors 1Bq *1IeIq oy O18! 10U YIIId)UD IT BnNvoag™ 61 “USI 3] GDP 1OUUNI It ‘UE 24) OME yori Mogi wos, Furyy 19Ad zopeym 14) ‘eatzord rou af og 4 Oje Aulpueypopun mnoyr~ oj 24 DIY ‘WIg) egw tpizy ap puy sr ‘ajqezed ay Fura swOd Tug pryse sapqioyip ain ‘ojdoed 24) w10I5 Roy 242 O} -ul pono sv ay uaym puy fT ‘NA qoTge eq 2y2 s9uIQOp 105 Fuiyovey ‘ou dius -1om Kays op. wies ur aqmop L *>ou aIOss sey 81 weIy J1249 Ng ‘sdiy Nays ata aw qaInowog adoad sys “warts 81 at se ¢saiti90dhy ‘nok yo paroydoid sereyg mie jam “msyy lun piry pur ‘posamjue aH 9 Qopuey usysraue tim persq ia sng ‘s12p|A 24q) 30 uonTpE!y 24) 02 dulpsooo8 saidia -yip Ay rou afew Aq “urig pose goqiirs pue saajtieqd 24) uly L § ‘safqui jo pure ‘ spegaa uajesq ‘ siod ae sdno jo dusgem ayise { ploy iar) * % MARE 16 And thefe ard they likewife which are fown on ftony ground, who, when they have heard the word, immediately receive it with felves, and fo endure but fora time, adtérward when affiGion, or perfecation arifeth for the words fake, immediatly they are offended, 18 And thefe are they which are fown among thorns: fuch as hear then the ear, after that the full corn in the ear 29 But ehenthe fruit is brought fori , immediatly he putteth inthe fhal we fiken the kingdom of God} or with what comparifon fhalwo compare it ? at Jr is like agrain of muftard feed: which, when iris fown imthe earth, is lefs than allthe feeds that prey pur ‘ino Jayiey aya few: “W)DAII[IQ 1 aye sduiu) [[e 3) ‘aity onun y "sn uo uowmeda op yues Bours 01 {S199 Mm 31) 0 WI Peo yey “PIIYD © JO ‘prey aured sty) aoun * yaysey siy fF *durotoy ‘pomo by parables, 4. ed avineyard , abour it, and Une wine-fat,.a 2 Anda husbandmen. a night receive-{ macn of the tru 3 And they. him, and feat Facsimile of part of No. 17. 2. Fic. The Scottish Catechism, with the Solemn League and Covenant. [See No. 1 in this list. ] 18. 80 Proceedings of the Royal Irish Academy. In the above notes I did not dwell sufficiently on the value of Patrick Neill’s advertisements, appearing at the end of some of the extant: works he printed, as evidence of his activity as a printer and the extent of the output of his press. In the third edition of his Catalogue, the late Mr. J. Anderson, at page 6, gave, in very brief form, two of Neill’s advertisements, containing about fifteen titles, not one of which was then (1890) known to be extant. About nine of these advertised titles have now been found and noted, some indeed imperfect or in a fragmentary state, but still there is thus confirmed the accuracy of Neill’s advertisements, and much encourage- ment is given to further search for those still untraced, and for perfect copies of those only at present evidenced by incomplete copies. I might add that a facsimile of the title-page of No. 9 (the Psalms) in this list will be found at page 14 of “Historical Memorials of the First Presbyterian Church of Belfast” (Belfast, 1887), and also of its silver-bound cover and clasps. Similarly, a reproduction of the title-page of the “New Testament in Verse,’ No. 8 in this list, will be found in Vol. XII of the “ Ulster Journal of Archaeology ” (1900), at page 41. (1) Sie) V. NOTES ON CERTAIN IRISH INSCRIPTIONS. By PROFESSOR R. A. 8. MACALISTER, Lirt.D. Puate VI. Read 24 January. Published 10 Aprir, 1916. During the summer of last year I had opportunity of examining a number of inscriptions, and made certain observations which I desire to bring before the Academy. As it happened, most of the inscriptions that came under my notice are extremely difficult to deal with. 1. KNOCKSHANWEE, Co. Cork. I have carefully re-examined this series of inscriptions, now making a brave show in the corridor of University College, Cork. Iam glad to be able to say that I have found nothing to modify in the readings contained in my paper on this important series of monuments.! I need only note that I am now certain that the name on the stone numbered 4 is CULRIGAI, and not the alternative there given, CUBBRIGAI. Moreover, | now doubt whether the inscription is imperfect after all. It seems at first sight to read, as I gave it in my previous paper, CULRIGAI MAQI MENU MAQI..., which implies that a name has been lost from the top of the stone; and while this reading is still possible, it may be that the last two words are really one, and that we are to read MENUMAQI as a name. 2. KILMARTRANN, Co. Cork. PuatE VI. About nine years ago I discovered an Ogham stone in a rath-cave on a townland called, on the O.S. map (6-inch, sheet 50), “Kilmartin Lower.” The real name of the place is, however, Kilmartrann, if we may trust local pronunciation. In the position of the stone it was impossible to read more than the first few letters, but these were such as to excite a lively desire to know how the inscription finished. I am now able to complete the inseription, having uncovered it with the invaluable co-operation of the Rev. Professor Power, of University College, Cork. The stone is a clay- 1 Proc. R.I.A., vol. xxxii, section C, no. 8. R,I.A. PROC., VOL. XXXIII., SECT. C, {13] 82 Proceedings of the Royal Irish Academy. slate, though more closely grained than many of the blocks of this — unsatisfactory material, which is so frequently used in Co. Cork for Ogham writing. It measures 5 feet 7 inches by 1 foot 8 inches by 10 inches. The inscription runs up the left angle, over the top, and a little way down the right angle. It is in excellent order; well and carefully cut in the first instance, every score is clear and fresh, except the A at the end of the first word, which has been broken off—probably by the rath-builders, when they adapted the stone as a lintel for the roof of their cave. The reading is as follows, beyond all possibility of doubt and dispute— UDDMENSA CELI NETTASLOGI. The last word is the easiest, so we begin by noticing that it evidently is the familiar name Nad-sluaigh, common in the genealogies and elsewhere in the as. literature, though not hitherto found in any form in Ogham. The second word marks the owner of the monument as a “ follower” or “tenant,” or in some such way a subordinate of this Nadsluaigh. In a previous paper’ I have enumerated the stones bearing this formula, and, as some may recollect, I have endeavoured to find it also hidden in the enigmatical inscription at Killeen Cormac. But the crux of the inscription lies in the first word, the name of this follower of Nadsluaigh. This name is absolutely unique in Ivish literature, so far as I, or the scholars that I have consulted, are able to say. The only ray of illumination, a feeble one at best, comes from the name UDDAMI, on one of the Whitefield stones; this does not help us much, as UDDAMI is itself highly problematical. Professor Mac Neill kindly allows me to quote the following ingenious note on the name, which he has sent me :— “Taking UDDMENSA to be genitive singular, to what declension is it to be referred? So far the only Ogham genitives ending in @ that have been identified belong to the consonant declension, a representing earlier asmés>mias, census>cis, sponsa>pés; later sensus>sians. The probability is that some vowel has disappeared between m and s, as in sinser<*senisser®<*senister®. So far, mt, genitive més, is the sole authenticated instance of a consonantal stem ending in s(<: the X letter at the end of the inscription, on the other hand, is formed of crossing lines, like an x. After the Q there follows immediately V..pDD...MxTY. The v might be an s—it is treated as S in Bishop Graves’ reading of the inscription—but the apparent last score is too indefinite to be taken as an intentional mark; it contrasts notably with the other three. After the v is a series of vowel-points, that look more like uv than anything else, and as such Brash has read them. There is room for a yowel-point before, and another after, the first of these apparent U’s, but I could not feel sure that they were actually there. If they were there, the combination of vowels would become 1u. Then come the two p’s, certain, though the scores of the first of these letters are damaged by a chip broken from the angle. After these come faint traces which formerly I read RA; I am now inclined to make them Gio. The B-surface is here quite worn and broken away, and covered with a thick growth of lichen; the tops of the scores of the apparent G are visible, and possibly it is not mere imagination that sees traces of the L under the lichen. If this growth could be removed, we might attain a more certain reading of this important inscription. Macanister—LWotes on Certuin Irish Inscriptions. 87 As to the verbation of the inscription, the beginning must be ANM CORRE, or ANM CORRK, according to the value to be assigned to the X character. But we are then confronted with an ambiguity. Are we to read MAQ V-UDDGLOMETT or MAQV— UDDGLOMETT? I incline to the latter. The combination of vowels IU in an Ogham word is rare, and UU unheard-of; though we must, on the other hand, postulate an unprecedented spelling for magi, namely, MAQVU or mMAQvi. If this be right, we have another name beginning with UDD-, but it is not any more intelligible than UDDAMI or UDDMENSA. Can these names be pre-Celtic ? Professor MacNeill, in the note printed above, has commented on the unusual collocation Ns in the latter name. The other stone from Aghabulloge, now in the museum of University College, Cork, shows this combination twice over. By most deplorable ill-luck this stone was used by masons as building material in the church of Aghabulloge, and they chipped away nearly all the H-surface, carrying off the H consonants and the vowel-points. By measure- ments of the tips of the B letters and of the spaces between them some approximation to a reading can be obtained; there is just room for ... NSaMa Netta aNSiLl AVI DETTAS, the capital letters denoting those characters which still exist in part, the minuscules those which can be inferred from the spaces. The L might bea G, but otherwise there is no doubt as to the reading of the surviving letters, and the restoration suggested is at least the most probable. 7. Knockoran, Co. Cork. A couple of years ago I examined this stone with the scholar whose recent death we all deplore, the late Sir John Rhys. We agreed in reading the latter part of the inscription as Brash had taken it, MAQI AILLUATTAN; but Sir John Rhys noticed an M before the opening word that had never been observed before, with a space after it that probably held five vowel-points. The inscription, therefore, begins MINNACCANNI, not ANNACCANNI. At my recent visit to Cork I confirmed this reading. The whole inscription is thus MINNACCANNI MAQI AILLUATTAN. No one looking at this stone with an unprejudiced eye can possibly doubt that the cross has been added to it at a date later than the inscription. 8. GLENNAWILLEN, Co. Cork. At the same time I corrected my previous copy of the two inscriptions on the Glennawillen stone, also in the College Museum. It is a very interesting case of the later appropriation of an Ogham stone for another inscription, unconnected with the first. The inscription on the left-hand edge is cut in 88 Proceedings of the Royal Irish Academy. very fine lines, and is difficult to read, especially towards the end, where the stone is much flaked. I now read this inscription COLOMAGNI AVI DUCURI. The other inscription, on the right-hand angle, is punched in broad and bold lines; it reads BRUSCO MAQI DOVAL