mode of fortification were the following: 1. The bad direction of the flank. After the introduction of ravelins and covered ways, the curtain became less and less the point of attack; the faces of the bastions now were chiefly assailed. To cover these well, the prolongation of the faces should have met the curtain at the very point where the flank of the next bastion was erected, and this flank should have been perpendicular or nearly so to this prolonged line (called the line of defence). In that case there would have been an effective flanking fire all along the ditch and front of the bastion. As it was, the line of defence was neither perpendicular to the flanks nor did it join the curtain at the curtain point; it intersected the curtain at 1, t, or of its length. Thus, the direct fire of the flank was more likely to injure the garrison of the opposite flank than the assailants of the next bastion. 2. There was an evident want of provision for a prolonged defence after the enceinte had been breached and successfully assaulted at one single point. 3. The small ravelins but imperfectly covered the curtains and flanks, and received but a poor flanking fire from them. 4. The great elevation of the rampart, which was all faced or revetted with masonry, exposed, in most cases, a height of 15 to 20 feet of masonry to the direct fire of the enemy, and of course this masonry was soon destroyed. We shall find that it took almost two centuries to eradicate this prejudice in favor of uncovered masonry, even after the Netherlands had proved its uselessness. The best engineers and authors belonging to the Italian school were: San Michele (died 1559), fortified Napoli di Romania in Greece, and Candia, and built Fort Lido near Venice; Tartaglia (about 1550); Alghisi da Carpi, Girolamo Maggi, and Giacomo Castriotto, who about the end of the 16th century all wrote on fortification. Paciotto of Urbino built the citadels of Turin and Antwerp (1560-'70). The later Italian authors on fortification, Marchi, Busca, Floriani, Rosetti, introduced many improvements, but none of these were original. They were mere plagiarists of more or less skill; they copied most of their devices from the German Daniel Speckle, and the remainder from the Netherlanders. They all belong to the 17th century, and were completely eclipsed by the rapid development of fortificatory science which at that time took place in Germany, the Netherlands, and France. The defects of the Italian system of fortification were soon discovered in Germany. The first man to point out the chief defect of the elder Italian school, the small bastions and long curtains, was a German engineer, Franz, who fortified for Charles V. the town of Antwerp. In the council held to try the plan, he insisted upon larger bastions and shorter curtains, but was outvoted by the duke of Alva and the other Spanish generals, who believed in nothing but the routine of the old Italian system. Other German fortresses were distinguished by the adoption of casemated galleries upon the principle of Dürer, as Küstrin,

fortified in 1537-'58, and Jülich, fortified a few years later by an engineer known under the name of Master John (Meister Johann). But the man who first broke completely through the fetters of the Italian school and laid down the principles on which the whole of the subsequent systems of bastionary fortification are founded, was Daniel Speckle, engineer to the town of Strasbourg (died 1589). His chief principles were: 1. That a fortress becomes stronger the more sides there are to the polygon which forms the enceinte, the different fronts being thereby enabled to give a better support to each other; consequently, the nearer the outline to be defended comes to a straight line, the better. This principle, demonstrated as an original discovery with a great show of mathematical learning by Cormontaigne, was thus very well known to Speckle 150 years earlier. 2. Acute-angled bastions are bad; so are obtuseangled; the salient angle should be a right one. Though correct in his opposition to acute salients (the smallest admissible salient angle is now generally fixed at 60°), the partiality of his time for right-angled salients made him hostile to the obtuse salient, which is indeed very advantageous and unavoidable in polygons with many sides. In fact, this appears to have been merely a concession to the prejudices of his time, for the diagrams of what he considers his strongest method of fortification all have obtuse-angled bastions. 3. The Italian bastions are far too small; a bastion must be large. Consequently, Speckle's bastions are larger than those of Cormontaigne. 4. Cavaliers are necessary in every bastion and on every curtain. This was a consequence of the system of siege of his time, in which high cavaliers in the trenches played a great part. But in Speckle's intention, the cavaliers were to do more than resist these; they are real coupures provided beforehand in the bastion, forming a second line of defence after the enceinte has been breached and stormed. The whole of the credit generally given to Vauban and Cormontaigne for cavaliers forming permanent coupures, is therefore in reality due to Speckle. 5. A portion, at least, of the flank, and better still the whole of the flank of a bastion, must be perpendicular to the line of defence, and the flank be erected in the point where the line of defence crosses the curtain. This important principle, the alleged discovery of which forms the greater part of the glory of the French engineer Pagan, was thus publicly proclaimed 70 years before Pagan. 6. Casemated galleries are necessary for the defence of the ditch; consequently Speckle has them both on the faces and flanks of the bastion, but only for infantry; if he had made them large enough for artillery, he would in this respect have been fully up to the latest improvements. 7. To be useful, the ravelin must be as large as possible; accordingly, Speckle's ravelin is the largest ever proposed. Now, Vauban's improvements upon Pagan consist partly, and Cormontaigne's improvements upon

Vauban consist almost entirely, in the successive enlargement of the ravelin; but Speckle's ravelin is a good deal larger than even Cormontaigne's. 8. The covered way is to be strength ened as much as possible. Speckle was the first to see the immense importance of the covered way, and he strengthened it accordingly. The crests of the glacis and of the counterscarp were formed en crémaillère (like the edge of a saw), so as to render enfilading fire ineffective. Cormontaigne, again, took up this idea of Speckle's; but he retained the traverses (short ramparts across the covered way against enfilading fire), which Speckle rejected. Modern engineers have generally come to the conclusion that Speckle's plan is better than Cormontaigne's. Speckle, beside, was the first to place artillery on the places of arms of the covered way. 9. No piece of masonry is to be exposed to the eye and direct fire of the enemy, so that his breaching batteries cannot be established before he has arrived on the crest of the glacis. This most important principle, though established by Speckle in the 16th century, was not generally adopted until Cormontaigne; even Vauban exposes a good deal of his masonry. (See C, fig. 2.) In this short abstract of Speckle's ideas the fundamental principles of all modern bastionary fortification are not only contained but plainly stated, and his system, which even now would afford very good defensive works, is truly wonderful considering the time in which he lived. There is not a celebrated engineer in the whole history of modern fortification who cannot be proved to have copied some of his best ideas from this great original source of bastionary defence. Speckle's practical engineering skill was shown in the construction of the fortresses of Ingolstadt, Schlettstadt, Hagenau, Ulm, Colmar, Basel, and Strasbourg, all of which were fortified under his direction.-About the same epoch, the struggle for the independence of the Netherlands gave rise to another school of fortification. The Dutch towns, whose old masonry walls could not be expected to resist a regular attack, had to be fortified against the Spaniards; there was, however, neither time nor money for the erection of the high masonry bastions and cavaliers of the Italian system. But the nature of the ground offered other resources in its low elevation above the water horizon, and consequently the Dutch, expert in canal and dike building, trusted to the water for their defence. Their system was the exact counterpart of the Italian: wide and shallow wet ditches, from 14 to 40 yards across; low ramparts without any masonry revetment, but covered by a still lower advanced rampart (fausse-braie) for the stronger defence of the ditch; numerous outworks in the ditch, such as ravelins, half moons (ravelins in front of the salient of the bastion), horn and crown works;* and finally,

A horn work is a bastionary front, two half bastions, a eurtain, and a ravelin advanced in front of the main ditch and closed on each side by a straight line of rampart and

a better use of the accidents of the ground than with the Italians. The first town fortified entirely by earthworks and wet ditches was Breda (1533). Subsequently the Dutch method received several improvements: a narrow zone of the scarp was revetted with masonry, as the wet ditches, when frozen over in winter, were easily passed by the enemy; locks and sluices were constructed in the ditch, so as to let the water in at the moment when the enemy had begun to sap the hitherto dry bottom; and finally, sluices and dikes were constructed for a systematic inundation of the country around the foot of the glacis. The writers on this elder Dutch method of fortification are Marolois (1627), Freitag (1630), Völker (1666), Melder (1670). An application of Speckle's maxims to the Dutch method was attempted by Scheither, Neubauer, Heidemann, and Heer (all from 1670 to 1690, and all of them Germans).-Of all the different schools of fortification, the French has enjoyed the greatest popularity; its maxims have found practical application in a greater number of still existing fortresses than those of all the other schools put together. Still, there is no school so poor in original ideas. There is neither a new work nor a new principle in the whole of the French school which is not borrowed from the Italians, the Dutch, or the Germans. But the great merit of the French is the reduction of the art to precise mathematical rules, the symmetrical arrangement of the proportions of the different lines, and the adaptation of the scientific theory to the varied conditions given by the locality to be fortified. Errard of Barle-Duc (1594), commonly called the father of French fortification, has no claim to the appellation; his flanks form an acute angle with the curtain, so as to be still more ineffective than those of the Italians. A more important name is Pagan (1645). He was the first to introduce in France, and to popularize, Speckle's principle that the flanks should be perpendicular to the lines of defence. His bastions are roomy; the proportions between the lengths of faces, flanks, and curtains are very good; the lines of defence are never longer than 240 yards, so that the whole of the ditch, but not the covered way, is within musket range from the flanks. His ravelin is larger than that of the Italians, and has a reduit or keep in its gorge, so as to admit of resistance when its rampart has already been taken. He covers the faces of the bastions with a narrow detached work in the ditch, called a counter-guard, a work which had already been used by the Dutch (the German Dillich appears to have first introduced it). His bastions have a double rampart on the faces, the second to ditch, which is aligned upon the faces of the bastions of the

enceinte so as to be completely flanked by their fire. A crown work consists of two such advanced fronts (one bastion flanked by two half bastions); a double crown work has three fronts. In all these works it is necessary that their rampart should be at least as much lower than that of the enceinte as the rampart of the ravelin to maintain the command of the enceinte over them. The adoption of such outworks, which of course were exceptions, was regulated by the nature of the ground.

serve as a coupure; but the ditch between the two ramparts is entirely without flanking fire. The man who made the French school the first in Europe was Vauban (1633-1707), marshal of France. Although his real military glory rests upon his two great inventions in the attack of fortresses (ricochet fire and parallels), still he is popularly better known as a constructor of them. What we have said of the French school is true of Vauban's method in the highest degree. We see in his constructions as great a variety of forms as is compatible with the bastionary system; but there is nothing original among them, much less any attempt to adopt other forms than the bastionary. But the arrangement of the details, the proportions of the lines, the profiles, and the adaptation of the theory to the ever-varying requirements of the locality, are so ingenious, that they appear perfection in comparison to the works of his predecessors, so that scientific and systematic fortification may be said to date from him. Vauban, however, did not write a line on his method of fortification, but from the great number of fortresses constructed by him the French engineers have tried to deduce the theoretical rules he followed, and thus have been established 3 methods, called Vauban's first, second, and third system. Fig. 1 gives the first system in its greatest simplicity. The chief dimensions were: the outer side of the polygon, from the point of one bastion to that of the next, 300 yards (on an average); on the middle of this line, a perpendicular a B, of the first; through B, the lines of defence from a" and a', a" d', and a'e". From the points a" and a', of a" a' measured on the lines of defence gives the faces a" c" and a' b'. From the shoulder points c" and b' arcs with the radius c' d' or b' e" were drawn between the lines of defence, giving the flanks b' d' and c"e". Draw e" d', the curtain. The ditch: with radius 30 yards, an arc in front of the point of the bastion, prolonged by tangents drawn to this arc from the shoulder points of the adjoining bastions, gives the counterscarp. The ravelin: from the curtain point e", with radius e" y (y, a point on the opposite face 11 yards beyond the shoulder-point), draw the arc yd, until it crosses the prolongation of the perpendicular a B; this gives the point of the ravelin; the chord to the arc just described gives the face, which is continued from the point until it reaches the prolongation of the tangent forming the counterscarp of the main ditch; the gorge of the ravelin is fixed by this line equally, so that the whole of the ditch remains free for the fire of the flanks. In front of the curtain, and there alone, Vauban retained the Dutch fausse-braie; this had already been done by the Italian Floriani before him, and the new work had been called tenaille (tenaglia). Its faces were in the direction of the lines of defence. The ditch in front of the ravelin was 24 yards wide, the counterscarp parallel to the faces of the ravelin, and the point rounded off. In this manner Vauban obtained roomy bas

tions, and kept his flanked salient angles well within musket range; but the simplicity of these bastions renders the defence of the place impossible as soon as the face of one bastion is breached. His flanks are not so good as Speckle's or Pagan's, forming an acute angle with the lines of defence; but he does away with the 2 and 3 tiers of uncovered guns which figure in most of the Italian and early French flanks, and which were never very useful. The tenaille is intended to strengthen the defence of the ditch by infantry fire, and to cover the curtain from direct breaching fire from the crest of the glacis; but this is very imperfectly done, as the breaching batteries in the reëntering place of arms (n, fig. 1) have a full view of the piece of the curtain next to the flank at e. This is a great weakness, as a breach there would turn all the coupures prepared in the bastion as a second line of defence. It arises from the ravelin being still too small. The covered way, constructed without crémaillères, but with trayerses, is much inferior to Speckle's; the traverses prevent not only the enemy, but also the defence, from enfilading the covered way. The communications between the different works are on the whole good, but still not sufficient for energetic sallies. The profiles are of a degree of strength which is still generally adopted. But Vauban still clung to the system of revetting the whole of the outside of the rampart with masonry, so that at least 15 feet high of masonry was uncovered. This mistake is made in many of Vauban's fortresses, and once made can only be remedied at an enormous expense by widening the ditch in front of the faces of the bastions, and constructing earthwork counterguards to cover the masonry. During the greater part of his life Vauban followed his first method; but after 1680 he introduced two other methods, having for their object to admit of a prolonged defence after the bastion was breached. For this purpose he took up an idea of Castriotto's, who had proposed to modernize the old tower and wall fortification by placing detached bastions, isolated, in the ditch, in front of the towers. Both Vauban's second and third methods agree in this. The ravelin is also made larger, the masonry is a little better covered; the towers are casemated, but badly; the fault that the curtain may be breached between bastion and tenaille is maintained, and renders the detached bastion partly illusory. Still, Vauban considered his second and third methods as very strong. When he handed over to Louis XIV. the plan for the fortification of Landau (second system), he said: "Sire, here is a place that all my art would not suffice to take." This did not prevent Landau from being taken 3 times during Vanban's life (1702, 1703, 1704), and again shortly after his death (1713).-The errors of Vauban were rectified by Cormontaigne, whose method may be considered as the perfection of the bastionary system. Cormontaigne (1696-1752) was a general of engineers. His larger bastions

permit the construction of permanent coupures and second lines of defence; his ravelins were nearly as large as those of Speckle, and fully covered that portion of the curtain which Vauban had left exposed. In polygons of 8 and more sides his ravelins were so far advanced that their fire took in the rear the besiegers' works against the next bastion as soon as he reached the crest of the glacis. In order to avoid this, two ravelins have to be conquered before one bastion can be breached. This mutual support of the large ravelins becomes more and more effective the more the line to be defended approaches a straight one. The reentering place of arms was strengthened by a reduit. The crest of the glacis is drawn en crémaillère, as with Speckle, but traverses are maintained. The profiles are very good, and the masonry is always covered by the earthworks in front. With Cormontaigne the French school closes, as far as the construction of bastionary defences, with outworks within the ditch, is concerned. A comparison of the gradual development of bastionary fortification from 1600 to 1750, and of its final results as laid down by Cormontaigne, with the principles of Speckle, as stated above, will tend to elucidate the wonderful genius of the German engineer; for although outworks in the ditch have been multiplied to an enormous degree, yet not a single important principle has been discovered during all these 150 years which had not been already clearly and distinctly enunciated by Speckle. -After Cormontaigne, the school of engineers of Mézières (about 1760) made some slight alterations in his system, the principal of which is the return to Speckle's old rule that the flanks must be perpendicular to the lines of defence. But the principal point for which the school of Mézières is remarkable is that they for the first time construct outworks beyond the covered way. On fronts particularly open to attack they place at the foot of the glacis, on the capital of the bastion, a detached ravelin called a lunette, and thereby approach for the first time to the modern system of permanent intrenched camps. In the beginning of the 19th century Bousmard, a French emigrant who served in Prussia and was killed at Dantzic in 1806, tried still to improve upon Cormontaigne; his ideas are rather complicated, and the most remarkable is that his ravelin, which is very large, is advanced to the foot of the glacis almost so as to take the place and functions, to a certain degree, of the lunette just described. A Dutch engineer of Vauban's time, who more than once opposed him in siege warfare with equal honor, Baron Coehorn, gave a further development to the old Dutch method of fortification. His system gives a stronger defence even than Cormontaigne's, by the clever combination of wet and dry ditches, the great facilities offered to sorties, the excellent communications between the works, and the ingenious reduits and coupures in his ravelins and bastions. Coehorn, a great ad

mirer of Speckle, is the only engineer of note who was honest enough to acknowledge how much he owed to him.-We have seen that even before the introduction of bastions, Albert Dürer used caponnières to afford a stronger flanking fire. In his fortified square he even entirely trusts to these caponnières for the defence of the ditch; there are no towers on the corner of the fort; it is a plain square with none but salient angles. To make the enceinte of a polygon entirely coincident with its outline, so as to have all salient and no reëntering angles, and to flank the ditch by caponnières, constitutes what is called polygonal fortification, and Dürer must be considered as its father. On the other hand, a star-shaped enceinte, in which salient and reëntering angles follow upon each other regularly, and in which each line is both flank and face at once, flanking the ditch of the next line with the portion next to the reentering angle, and commanding the field with the portion next the salient-such an outline constitutes tenaille fortification. The older Italians and several of the older Germans had proposed this form, but it was not developed till afterward. The system of George Rimpler (engineer to the emperor of Germany, killed in defending Vienna against the Turks in 1683) forms a kind of intermediate stage between the bastionary and tenaille system. What he calls intermediate bastions constitute in reality a perfect line of tenailles. He declared himself energetically against open batteries with a mere earth parapet in front, and insisted on casemated batteries wherever they could be erected; especially on the flanks, where 2 or 3 tiers of well covered guns would thus have a far greater effect than the 2 or 3 tiers of guns in open flank batteries, which could never act together. He also insisted on batteries, that is, reduits, in the places of arms of the covered way, which Coehorn and Cormontaigne adopted, and especially a double and triple line of defence behind the salient angles of the enceinte. In this manner his system is remarkably in advance of his time; the whole of his enceinte consists of independent forts, each of which has to be taken separately, and large defensive casemates are used in a manner which reminds us, almost in the details even of their application, of the more recent constructions in Germany. There is no doubt that Montalembert owed as much to Rimpler as the bastionary system of the 17th and 18th century to Speckle. The author who first fully developed the advantages of the tenaille over the bastionary system was Landsberg (1712); but it would lead us too far if we were to enter into his arguments or describe his fortificatory outline. Of the long series of skilful German engineers who followed Rimpler and Landsberg, we may name the Mecklenburg colonel Buggenhagen (1720), the inventor of blockhouse traverses, or traverses hollowed out and adapted for casemated musketry fire; and the Würtemberg major Herbort (1734), inventor of defensive barracks,

large barracks in the gorge of salient works, proof against vertical fire, with embrasured casemates on the side facing the enceinte, and barracks and store rooms on the side facing the town. Both these constructions are now very largely used. Thus we see that the German school, with almost the only exception of Speckle, was from its origin adverse to bastions, which it sought to replace chiefly by tenailles, and that it attempted at the same time to introduce a better system of inner defence, chiefly by the use of casemated galleries, which again were considered as the height of absurdity by French engineering authorities. One of the greatest engineers, however, that France ever produced, the marquis de Montalembert (171399), major-general of cavalry, passed over with drums beating and colors flying into the camp of the German school, to the great horror of the whole French engineering corps, who, up to the present date, decry every word he has written. Montalembert severely criticized the defects of the bastionary system; the ineffectuality of its flanking fire; the almost certainty it offered to the enemy that his shots if they missed one line must do harm in another; the want of protection against vertical fire; the perfect uselessness of the curtain as to fire; the impossibility of having good and large coupures in the gorges of the bastions, proved by the fact that no fortress of his time had any of the multifarious permanent coupures proposed by the theorists of the school; and the weakness, bad connection, and want of mutual support of the outworks. Montalembert therefore preferred either the tenaille or the polygonal system. In either case the body of the place consisted of a row of casemates, with one or two tiers of guns, the masonry of which was covered from direct fire by a counterguard or couvre-face of earthwork extending all around and having a second ditch in its front; this ditch was flanked by casemates in the reëntering angles of the couvre-face covered by the parapet of the reduit or lunette in the reëntering place of arms. The whole system was based upon the principle of opposing, by means of casemated guns, such an overwhelming fire to the enemy the moment he reached the crest of the glacis, or of the couvreface, that he could not possibly succeed in erecting his breaching batteries. That casemates could do this he maintained against the unanimous condemnation of French engineers, and he afterward even compiled systems of circular and tenaille fortifications in which all earthworks were rejected and the whole defence intrusted to high casemated batteries with from 4 to 5 tiers of guns, the masonry of which was to be protected by the fire of its batteries only. Thus, in his circular system, he contrives to concentrate 348 guns on any point 500 yards from the fortress, and expects that such an immense superiority of fire would put the possibility of erecting siege batteries entirely out of the question. In this, however, he has found no adherents, except in

the construction of the sea fronts of coast forts; here the impossibility of breaching strong casemated walls by the guns of ships was pretty well demonstrated by the bombardment of Sebastopol. The splendid forts of Sebastopol, Cronstadt, Cherbourg, and the new batteries on the entrance of Portsmouth harbor (England), and almost all modern forts for harbor defence against fleets, are constructed according to Montalembert's principle. The partly uncovered masonry of the Maximilian towers at Lintz (Austria) and of the reduits of the detached forts of Cologne are imitated from Montalembert's less happy projects. In the fortification of steep heights (Ehrenbreitstein in Prussia, for instance) the uncovered masonry forts have also been sometimes adopted, but what resistance they will be able to make must be decided by actual experience. The tenaille system has never, to our knowledge at least, found practical application, but the polygonal system is in great favor in Germany, and has been applied to most modern constructions there; while the French tenaciously cling to Cormontaigne's bastions. The enceinte, in the polygonal system, is generally a plain earthwork rampart with revetted scarp and counterscarp, with large caponnières in the middle of the fonts, and with large defensive barracks behind the rampart and covered by it to serve as coupures. Similar defensive barracks have also been erected as coupures in many bastionary works, to close the gorges of the bastions; the rampart serving as a counterguard to protect the masonry from distant fire. Of all Montalembert's proposals, however, that of detached forts has had the greatest success, and initiated a new era, not only in fortification, but in the attack and defence of fortresses, and even in general strategy. Montalembert proposed to surround large fortresses in important situations by a single or double chain of small forts, on commanding elevations, which, though isolated in appearance, would still support each other by their fire, and, by the facility they gave for large sorties, would render a bombardment of the place impossible, and when required form an intrenched camp for an army. Vauban had already introduced permanent intrenched camps under the guns of fortresses, but their intrenchments consisted of long continuous lines, which, if broken through at one point only, were completely at the mercy of the enemy. But these intrenched camps of Montalembert's were capable of a far greater resistance, for each fort had to be taken singly, and before 3 or 4 at least were conquered, no enemy could open his trenches against the place. Moreover, the siege of each of the forts could be interrupted at every moment by the garrison, or rather the army encamping behind the forts, and thus a combination of active campaigning and regular fortress warfare was secured, which must greatly strengthen the defence. When Napoleon led his armies hundreds of miles through the enemy's country, never heeding the fortresses which had all been constructed