The strong suppression of higher frequency (HF) wind-waves by lower frequency (LF) waves has been observed in laboratory experiments since 1966 (e.g. Mitsuyasu, 1966), but theories explaining the physical cause have varied fundamentally, all of which have aspects which may be challenged. As a result, there is no consensus as to which theory correctly describes the mechanism. This thesis presents a study which has been undertaken to determine the fundamental nature to the mechanism, through detailed measurement of the interaction between HF and LF waves. These measurements were primarily made in a laboratory wind-wave flume, with wind and paddle waves propagating in the same direction, and an array of up to 24 wave gauges distributed throughout the flume. An experiment was also performed in natural conditions in a coastal lake, with a point wave gauge providing wave records, and multiple sonic anemometers providing 3D turbulent wind velocity data.The temporal transition in the laboratory from conditions with pure wind-waves to conditions of wind + paddle forcing has been examined in detail, through repeating a specific transition sequence 186 times. The aggregation of these measurements provided a clear representation of the precise timescale of HF wave suppression, revealing that the majority of suppression occurs too quickly to have been caused by reduced wind-input, instead indicating a mechanism of enhanced HF wave dissipation. The examination of spatial variation in HF wave energy along the LF wave phase indicated that the majority of suppression occurs on the LF wave crest, and high on the windward (rear) face, which locations are prone to experience the highest wind velocities along the LF wave phase. This observation also argued against suppression primarily by a mechanism of reduced wind-input. Quantification of HF wave suppression versus a broad range of wind velocities, paddle wave conditions and fetches did not reveal any critical dependence of suppression on either LF wave breaking or wind separation at LF wave crests, due to strong suppression occurring when neither of these was occurring. It was concluded that suppression occurs primarily as a result of enhanced dissipation of HF waves near LF wave crests, and possibly to a lesser degree from reduced wind input, the latter likely caused by wind separation at LF wave crests. The enhanced dissipation of HF waves appears correlated with locations of strong wind input, suggesting that the increased wind velocity near LF wave crests exceeds the level of forcing which HF waves of low are able to withstand without breaking. It is concluded that HF waves experience both an excess wind forcing near the LF wave crest, and a deficit of forcing in the LF wave trough, both of which result in less HF wave energy than would be present in absence of LF waves.In conditions of low paddle wave steepness, the distribution of energy input between LF and HF waves, and the LF waves’ capacity to absorb wind input, exhibited substantial nonlinearity versus LF wave steepness and wind velocity. LF paddle waves grew negligibly below a critical steepness and wind velocity, after which they ‘caught the wind’, absorbing the majority of wind input, and grew rapidly. This suggests that a change in the wind flow regime occurs at the critical steepness and wind velocity, mostly likely due to wind flow separation at LF wave crests (Donelan et al., 2006). These observations suggest that the Jeffreys 1924 mechanism may be more quantitatively significant to wave growth than the commonly applied theory of Miles (1957), and also provided further argument that wind separation contributes to HF wave suppression on the leeward face and in troughs of LF waves, although secondary in magnitude and slower acting than the enhanced HF wave dissipation at and windward of LF wave crests.HF wave growth was unexpectedly observed to be enhanced by LF waves of very low steepness, low wind velocity and short fetch. Much of this HF wave energy increase occurred at harmonic frequencies of the paddle wave, yet without being bound to the paddle wave, which is not explained by current wave theories. This suggests that LF waves and their harmonics may provide seeding to assist incipient HF waves to reach a point of Kelvin-Helmholst instability, or ripple formation, earlier. The marginally accelerated wind at LF wave crests of even very low steepness may also cause this threshold to be reached earlier.The statement by Plant and Wright (1977), that the longest waves to grow primarily by wind input are ca. 0.1m in length, with waves longer than this growing primarily by nonlinear interactions, was challenged by measurements made in the present study. A several-fold increase in the height of wind-forced monochromatic LF waves with wavelengths of order 1m, in conditions with minimal HF wave energy from which energy could be transferred, indicated that these waves much longer than 0.1m grew as a result of wind input.The results of the coastal lake experiment were inconclusive, with no clear patterns observed which could prove suppression to be present in nature. The range of conditions measured was limited, suggesting further insight may be gained by future experimentation with a broader range of conditions