Katana vs longsword: curve, edge, guard, and armor

A katana and a longsword are both swords you can hold in two hands. A katana is a curved Japanese blade with one edge, worn edge-up at the waist. A longsword is a European weapon with a straight or lightly tapered blade, usually two edges, and a bar across the hilt. That bar is the crossguard. Its arms are quillons. The katana guard is a plate, the tsuba, sitting in front of the tsuka, the wrapped grip. Each sword was built to meet a different hold and a different shell of armor. This page stays with the swords. The people who carried them are on the samurai and knight page.

Weights below are named objects, not a blended average. The Met katana records used here give length and curvature and do not give weight, so there is no average katana weight on this page. One Royal Armouries blade, XXVIS.331, weighs 0.72 kg by itself. When a catalog lists a complete longsword, that weight includes the hilt and pommel. It is not the same kind of measurement as a bare blade.

Side by side

Read the last column as a change in the hand, not as a score. Dates on these objects run from the 15th century into the 17th. A sword of 1420 and a sword of 1620 are not a pair that ever had to meet.

Curve, one edge, two edges

The Metropolitan Museum's catalog line for a katana is plain: a curved blade longer than 60 cm, in an uchigatana mounting, worn in a waist sash with the cutting edge facing up. On blade 36.25.1669 the cutting edge is 70.2 cm and the curvature is 1.8 cm. A 15th century blade, 36.25.1660, has a cutting edge of 73.6 cm and a curvature of 2.6 cm. A late 17th century Osaka blade, 36.25.1658, has 72.4 cm of edge and only 0.96 cm of curve. Masazane's blade dated 1526 (2001.574) has a cutting edge of 75.1 cm. A mounted sword, 91.2.67, has a cutting edge of 60.1 cm and a curvature of 1.4 cm. Royal Armouries XXVIS.331, signed by Masanori and Masakatsu and dated by the museum to the 17th century, has a nagasa of 694 mm and a sori of 9 mm. On this set the edge runs about 60 to 75 cm, and the bend runs about 0.9 to 2.6 cm. Poster art likes a deep crescent. Several of these blades barely leave a straight line.

A curve changes the cut because the edge does not hit and bounce off in one smack. It meets, then keeps moving along the target while the hands finish the stroke. That slide is a draw cut. The bend helps the edge stay in the cut for that slide. A straight edge can draw too, if the hands pull it, but the geometry favors a chop or a thrust. The longsword point sits on the center line of the blade, so a shove from the back hand drives it. The Met says this outright about sword 2006.564, probably German, about 1400 to 1430: a steep point and a stiff blade, "primarily intended for thrusting," sturdy enough "to pierce armor." That sword is 124.8 cm overall, with a blade of 97.8 cm, and it weighs 3 pounds 7 ounces (1,560 g).

Two edges change the thrust and the bind. A bind is the moment the blades are crossed and both people are pressing. On a longsword, either edge can be the sharp side of that press, and the point is still aimed between them. You can cut on the way out and on the way back without flipping the sword over. A katana has one edge, the ha, and a back, the mune, that does not cut. In a bind you have to keep the edge where you want the cut. The back can take a shove. It will not bite the way the second European edge will.

A fuller is the groove forged or ground along a blade. Dirk H. Breiding, in the Met's 2004 essay on arms and armor myths, says the groove lightens the blade without weakening it or taking away its flexibility. He rejects the name "blood groove." Royal Armouries IX.2149, the 15th century hand-and-a-half sword Oakeshott published as Type XVIIIa, is a flattened diamond with no need of a deep groove to make the point stiff. Other European blades carry a shallow fuller for a long stretch of the edge. The groove is a choice about mass, not a secret about bleeding.

1. The katana sits edge-up in the sash, which is the Met's wear, not a movie back-draw.
2. The draw brings that one edge out across whatever is in front of the scabbard mouth.
3. The curve keeps a length of edge in the cut while the hands finish the stroke.
4. The point is still there, but it is not lined up the way the stiff point on Met 2006.564 is lined up for a thrust.

Folded steel and bloomery steel

Hunter starts the Japanese material at tama-hagane, steel from iron-rich sand smelted in a tatara. He is blunt about it: the steel is full of impurities, and the carbon is not spread evenly. Too little carbon and the metal is soft. Too much and it is brittle. Kitae, the fold, is how the smith answers that. Pieces are forge-welded into a block, hammered out, and folded back. Impurities work out. Carbon evens out. The Met's own example is that fourteen folds make over 16,000 layers. That is an illustration of the arithmetic, not a rule that every smith stopped at fourteen. The folds also make the jihada, the surface grain. Masame is a straight grain. Ayasugihada is a curved one. Masazane's 1526 blade is the Met's rare case of ayasugihada outside the Gasan school.

The essay then wraps that harder outer steel, kawagane, around a softer iron core, shingane. The skin holds an edge. The core is there to resist a break. Other Japanese constructions exist. This is the one the Met essay teaches, and it matches the quench: hard where you cut, tougher where the blade has to survive the shock.

European sword iron of the early Middle Ages was bloomery iron. Alan Williams, then archaeometallurgist at the Wallace Collection, put it this way in a 2009 Gladius paper on Viking-age blades: ore and charcoal in a small furnace, the iron never fully melted (pure iron melts at 1550°C), slag still in the bloom. Some of that iron took up carbon and became steel. Sorting the steel out on purpose was, in his words, extremely difficult. Quenching a 0.8% carbon steel could jump hardness from about 250 VPH into the 600 to 800 range, and it also made the piece brittle. A gentle reheating, the temper, was supposed to trade some of that hardness for toughness so a blade could flex instead of chip. Williams says the control was poor, and heat treatment was frequently skipped.

Those Viking blades are earlier than the longsword on this page. The Oakeshott Institute's 18 October 2016 note looks at hardness tests through the 16th century and still refuses a single spring-steel story. Most tested period swords were a steely iron mix. Carbon and hardness jump inside one blade. In their 14th to 16th century group, seven blades ran from 136 VPH to 650 VPH, with piled, billet, core-wrapped, and carburised construction, and with both slack quenches and quench-and-temper work. A possibly 12th century sword in the Wallace Collection (A457), tested by David Edge and Williams at 5 cm intervals, averaged 211 VPH in the middle and 279 VPH along the edge. That is a harder edge than a core, which rhymes with the Japanese idea, and the numbers are far below a modern spring blade held at one Rockwell target.

The practical split is this. Differential hardening, when the clay quench works, gives a katana a hard edge and a body that is tougher than that edge. A European quench and temper, when it works, pulls brittleness back through the section so the blade can bend and spring instead of staying glass-hard. The Institute's point is that medieval steel rarely gave the smith a uniform section to temper. Modern replica steels they list, such as 5160 (about 0.57 to 0.62% carbon) and 6150 (about 0.48 to 0.53%), are mixed to a narrow window. They call homogenous mill steel the largest leap from the original material. If a new longsword flexes and pings back in your hands, you are often feeling that modern steel, not a promise about a river-found blade of 1420.

Tsuba, tsuka, and the cross

The tsuka is the grip: wood over the tang, ray skin, silk cord. Both hands live on it, and both stay on the hilt side of the tsuba. The plate stops the hand running onto the edge in a cut, and it is also a small plate in front of the knuckles. It does not grow side rings. A blade that slides in from the side can still find fingers. The Met lists tsuba with the rest of the mounting, next to menuki and a ray-skin grip, as metalwork around the blade rather than part of the heat treatment.

A European hilt, in the Met Journal account of sword pommels, is three parts on one tang: guard, grip, pommel. The tang is the narrow unsharpened shank forged in one piece with the blade. The guard slides on until it sits against the shoulders, then the grip, then the pommel. The tang tip is peened flat so the stack cannot walk off. The pommel keeps the hand from leaving the end, and it counterweights the blade. On 2006.564 the pommel is engraved MARIA, a pious invocation to the Virgin, and the crossguard has a slight reverse curve.

• Quillons reach sideways and a little forward. A blade skidding down your edge hits the bar instead of the hand.
• The same bar can hook, pin, or strike at close range. The Royal Armouries say swords of IX.16's type used the crossguard and pommel for wrestling and striking.
• Half-swording puts one hand on the blade. That hand is past the guard. You accept the risk because the point, choked up short, can be pushed into a gap.
• Side rings show up on later complex hilts. A Royal Armouries swept-hilt rapier, IX.109, is a different object from the plain cross on IX.2149.

Hand protection is the practical difference. The tsuba guards the slide forward and leaves the sides open. The crossguard guards the slide down the edge and gives you a lever. A later basket covers more of the hand than either of these. You fight with the holes your guard actually leaves.

Draw cut, bind, and the hand on the blade

The katana's home is the waist. Edge-up in the sash means the draw and the cut can be one motion. That is the wear kenjutsu inherits, and it is why curvature is tuned for a drawing slice rather than for a couched lance. The katana page is the place for mountings, the short companion sword, and the long argument about spears and guns on a battlefield. This page only needs the design brief: a sidearm you wear, draw, and cut with, two hands on the tsuka, one edge doing the work.

The longsword is an arming weapon. You carry it when you are already equipped to fight, on foot or from the saddle, and you use it from the bind, from the thrust, and, in armor, with a hand on the blade. Met 14.25.1196, 15th century, European or possibly British, is the clean definition of the grip: the hilt is long enough for one hand or two. It is 122.9 cm overall, blade 93.7 cm, and it weighs 3 pounds 9 ounces (1,616 g). The Royal Armouries description of IX.16, an early 15th century longsword they associate with the kind of fighting at Agincourt in 1415, says the stiff acute point was for half-swording into gaps such as the mail at an armpit. IX.1106, a mid-14th century sword from Lake Constance, gets the same note: stiff, heavy in the blade, specialized for armored fighting and for gripping the blade with one hand. Those notes are the museum talking about use. They do not say every longsword was built to stab plate.

Ewart Oakeshott's typology, as the Oakeshott Institute summarises it, sorts medieval swords by blade and hilt, from The Sword in the Age of Chivalry (1964, revised 1981) and Records of the Medieval Sword(1991). The Royal Armouries call IX.2149 Type XVIIIa in that 1991 book: straight, two-edged, flattened diamond section, pear pommel, straight quillons. Oakeshott's revised date in the Armouries note is about 1400 to 1440. Overall length 1,070 mm, blade 833 mm and slightly shortened, weight 3 pounds 6 ounces. Put that next to the two Met hand-and-a-half swords and you have a cluster, 3 pounds 6 ounces to 3 pounds 9 ounces, for three named weapons. It is not the average of every longsword ever forged.

Names beginners swap

The arguments get messy when four swords are called by one word. Length overlaps. A tachi cutting edge in the Met can sit in the same 60 to 75 cm band as a katana. Mounting is the split the museum actually prints on the label.

A tachi in the Met, object 10.187.26, has a 16th century blade and a mid-19th century mounting: cutting edge 69.8 cm, curvature 2.65 cm, worn edge-down. Another, 36.25.1679, late 12th to early 13th century, has a cutting edge of 64.3 cm. The steel can look like a katana's steel. The cord and the edge-down hang are the tachi. If a label says "slung sword (tachi)," do not correct it to katana because the blade is curved.

On the European side, the single-handed sword is what English collectors call an arming sword. Breiding's 2 to 4 pound average belongs there. The longsword adds grip length so the second hand can join, or so one hand can leave and take the blade. The greatsword, or two-handed sword of war, is the long one. The Met's British or Western European two-handed sword of about 1400 to 1450 weighs 5 pounds 3 ounces (2,354 g). Breiding's ceiling for that group, rarely above 10 pounds, is the number to set beside a game stat.

Plate, lacing, and which blow matters

Armor picks the blow. Stuart Pyhrr, on the Met's 14th century yoroi associated with Ashikaga Takauji (14.100.121), contrasts it with European plate in one sentence of shop talk: European armor is formed of steel plates, and this yoroi is small plates of iron and leather, lacquered, laced in silk. Overlaps make a double thickness. The lacing lets the skirt telescope when a man sits on a horse. The Museum dates that yoroi type from about the 10th century to the 14th. A slice has cords, edges of plates, and gaps at the joints to meet. A solid breastplate does not offer the same catch.

Breiding dates a developed plate harness to about 1420 to 1430. A field armor, he writes, usually weighs 45 to 55 pounds (20 to 25 kg), with the weight spread over the body, each limb in its own plates and lames. That shell is why a stiff point mattered. Piercing armor, in the Met's words on 2006.564, is the job. Mail still sat in the openings. The Royal Armouries note on IX.16 points the thrust at those openings, not at the middle of a cuirass. Breiding also says a skilled sword could sever a limb and cut through armor. Read that beside the pierce-armor sentence, not instead of it. Some armor yields to a cut. A fitted plate breastplate is the case the longsword point was drawn for.

Later Japanese armor moved toward larger plates. The samurai armor page carries that change. The warrior comparison, including what each fighter actually carried besides a sword, stays on samurai vs knight. For this page the short version is enough. Lamellar and lace make the drawing cut a sensible design. Plate makes the stiff thrust and the hand on the blade a sensible design. Change the armor and the blow that matters changes with it.

FAQs

One measured curve

Royal Armouries IX.2149 is a straight two-edged Type XVIIIa blade that the same record now describes as slightly curved in the plane of the blade, with a river patina. That bend is damage on a shortened edge of 833 mm. Katana XXVIS.331 has a sori of 9 mm on a nagasa of 694 mm, recorded as slight koshi-zori. Two bends, two causes, both small enough to miss if you only know the swords from a poster.