The above image shows the tracks of all the tropical cyclones during the 2017 Atlantic hurricane season. Despite this season being extremely active and devastating, there is a "hole" in the map over the Caribbean, especially the eastern portion. No cyclones developed in 2017 in the eastern Caribbean and the ones that entered this region died out (though Harvey ultimately regenerated further west, with horrific consequences). This particular season was cherrypicked, but this phenomenon is sufficiently common and well-known to the meteorological community that the eastern Caribbean is sometimes known as the "Hurricane Graveyard".
Statistics confirm that the Hurricane Graveyard is more than just weather lore. Consider the diagram below, which shows the point of genesis for all known "in-season" (June to November) tropical cyclones between 1851 and 2008:
It is clear that, relative to other areas at the same latitude, the eastern Caribbean churns out far fewer tropical cyclones. What is more, tropical cyclones often unexpectedly weaken or dissipate as they pass through it, contrary to the expectations of many computer models and forecasts (however, models do fairly well at predicting the scarcity of genesis in the eastern Caribbean, see this article). So why does the Hurricane Graveyard exist?
A 2010 study examined some the factors that contribute to this phenomenon. One of these is the Caribbean low-level jet (CLLJ). Low-level jets are channels of fast-moving air in the low levels of the atmosphere near the surface (analogous to jet streams, which are much stronger and at higher altitudes). The CLLJ blows east to west over the Caribbean, with the strongest winds occurring near its center, especially during July when the CLLJ reaches its peak. The winds moderate and reach a relative minimum by October, coinciding tellingly with a somewhat greater rate of tropical cyclone formation in the east Caribbean.
The above image shows the low-level winds in the Caribbean region averaged over the month July 2006. As the diagram indicates, wind speeds in excess of 10 m/s (22 mph) are typical during that time of year. These winds are due in part to the Azores-Bermuda high pressure system, a persistent area of dry, sinking air over the central subtropical Atlantic around which air flows clockwise. The southwestern portion of clockwise flow is evident in the diagram above. But why do horizontal winds of this sort inhibit tropical cyclone development? Further, the peak winds occur toward the center of the Caribbean, whereas our area of interest lies to the east, so what gives?
The key concept explaining the hurricane graveyard is low-level divergence. Divergence is simply a measure of to what degree air (or any fluid) is flowing out of a region. Two schematic vertical cross-sections of the atmosphere are shown below.
In the first, there is low-level divergence, as air is flowing outward from the central region. Since any outgoing air must be replaced, this air can only come from above, creating downdrafts and convergence (the opposite) at the atmosphere's upper levels. The reverse occurs in the second diagram. One may note that our diagram of Caribbean winds does not show air moving outward from any central point. However, the CLLJ accelerates moving east to west across the eastern portion of the sea, so it is still a situation where more air is leaving this region in the lower levels than is entering it. Hence the eastern Caribbean exhibits low-level divergence!
Crucially, low-level divergence is the opposite of what hurricanes need to thrive. Thunderstorm activity is driven by rising moist air and killed by sinking dry air. The divergence caused by the CCLJ often causes the thunderstorms associated with tropical cyclones or disturbances to collapse when they pass over the region, at last accounting for the Hurricane Graveyard.
Recent analyses have shed even more light on this phenomenon. It peaks in July when the CCLJ does, diminishing toward the end of hurricane season. Indeed, the rate of cyclonogenesis in the eastern Caribbean more closely resembles that of the rest of the basin from October onward. In addition, the Hurricane Graveyard effect is more pronounced in summers with an El Niño, when the Azores-Bermuda high is farther west and stronger, leading to a more powerful CCLJ. This effect illustrates how there is a great deal more to hurricane formation than humid air and warm water. Several other atmospheric factors, including low-level divergence, feature crucially in understanding and forecasting the development of tropical cyclones.
Sources: https://journals.ametsoc.org/doi/abs/10.1175/2009BAMS2822.1, https://journals.ametsoc.org/doi/full/10.1175/WAF-D-13-00008.1, https://journals.ametsoc.org/doi/pdf/10.1175/2011JCLI4176.1, http://www.atmo.arizona.edu/students/courselinks/fall16/atmo336/lectures/sec1/winds_fall15.html