Horizontal clutch is the oldest of the two engagement systems used in mechanical chronographs, and it still divides designers today. Its principle is simple: a toothed wheel, flat relative to the movement, slides laterally in the plane of the caliber to engage the chronograph's seconds wheel. When you press the start button, engagement occurs in the horizontal plane, hence the name.
The historical problem with this system is stuttering: that small visible jump on the seconds hand at the moment of engagement. It happens because the teeth of the two wheels are not in phase when they touch. It's not a manufacturing defect, it's a geometric consequence of lateral engagement. You can see it clearly on certain vintage calibers, the Lemania 2310 leading the way, and it never entirely disappears. Agenhor patented a micro-toothed variant that reduces the phenomenon appreciably, but doesn't eliminate it. F.P. Journe got around the problem differently, with an oscillating pinion that changes the engagement logic at its root.
Why then do manufacturers continue to use it? Three reasons. First: it's visually more legible from above, which matters on skeletonized movements or with transparent casebacks. Second: it integrates well with the column wheel, the control system it pairs with in the El Primero caliber, in the Seagull ST-1901B, in the Czapek SXH6 and in dozens of others. Third: in flyback version, like in H. Moser's Calibre HMC 700, it can work on two successive stages allowing rapid and controlled transitions that vertical clutch achieves with greater mechanical difficulty.
The vertical clutch, by comparison, engages along the movement's axis: the surfaces touch face-to-face like two friction discs, which eliminates stuttering and guarantees linear engagement. Purists prefer it. But it takes up more vertical space, complicates reading the movement from outside, and in ultra-thin calibers it becomes a serious problem. I've had on the bench calibers where designers chose horizontal clutch precisely to gain tenths of a millimeter of thickness.
On pocket chronographs from the early twentieth century, horizontal clutch was almost universal, paired with a cam instead of a column wheel. The cam is less precise but cheaper to produce: Lémania was still using it in 1877, as in the Franck Muller Endurance 24. The column wheel gradually gained the upper hand in the high-end segment, but the cam survives in mid-range movements.
From the bench: when I disassemble a chronograph with horizontal clutch and see unusual wear marks on the engagement teeth, the problem is almost never the clutch itself but incorrect reassembly that left excessive play between the two wheels. Engagement distance is as critical as surface quality.