Agreeing with Nigel, a maximum gain is not achieved when the antenna is one wavelength long. In fact, gain will continue to increase as length increases, but in a complicated (ie. not convenient and often not useful) way. The pattern breaks up into a series of lobes which are too difficult to describe here, but I'll try anyway. As the length of your wire increases beyond one wavelength, the radiation pattern changes from squished ball to a sort of a peanut shape. And as you go longer still, the peanut turns into a couple of lumpy cones joined at their pointy ends. These cones become narrower and narrower as your antenna increases in length. But even these descriptions depend on the location of your feedpoint with the simplest being in the center of the antenna. But as Nigel points out, this is the worst place to feed an antenna that is exactly an integer multiple of one wavelength since the impedance is so very very high that arranging an impedance matching network to affect maximum power transfer from feedline to antenna becomes unworkable.
The principle that the pattern gets narrower and has higher gain in very specific directions is behind various long wire antennas including the Rhombic which is usually many wavelengths long, but is arranged so that the cones of maximum gain of two longwire sections combine constructively along the axis of the rhombic to from a very high gain antenna indeed but one with a very narrow beamwidth.
This concept of the directivity breaking into cone shapes is hard to explain here, so it may be time to crack open a text book or try it out on a simulator if you want to see a picture of the patterns. Such beamforming using longer antennas is often very inconvenient, but justified in some cases. The trend in the industry in recent years has been in the opposite direction, that is towards getting good performance from electrically small antennas of less than half wavelength, rather than going larger. This is driven mainly by the demand for ever smaller portable devices with integral wireless functions that the industry is dealing with now.
The best place to see examples of electrically large antennas, like multiple wavelength wires, remains in the services using LF, MF, HF and low VHF frequencies, as once you go above 100 Mhz or so, most people do their beamforming with driven or parastic arrays instead.
The term wavelength is useful in many ways, not just in describing the length of wire antennas.