Absolute blood volumes were determined in animals by means of bleeding in the late 1800s. This approach was for obvious reasons not applied to humans. The use of carbon monoxide (CO) as a tracer to determine blood volume was first proposed by French scientists Grehant and Quinquaud in 1882. Soon after a usable set-up for human use was developed by eminent Oxford scientists John Haldane and Lorrain Smith and presented in the Journal of Physiology in 1900 (in which they also report on the O2 content of arterial blood).
With the aim to ensure that all of the inhaled CO would also bind to hemoglobin and become distributed equally throughout the circulation, Haldane and Smith performed the measurements using an impressive 2 hours of re-breathing. During these early days of CO re-breathing Haldane would later report that studies were finally abandoned when the soda lime (used to eliminate CO2 in the rebreathing circuit) would make its ways through the fabric of the re-breathing rubber bag, something that fortunately does not occur today.
Despite these and other technical limitations, their measurement precision was an impressive ~8%. Haldane and his Oxford colleague Claude Douglas (later famous for among others the Douglas bag method) applied their CO re-breathing method to study the effects of altitude exposure on hemoglobin mass during research expeditions to Mount Teide on Tenerife in 1910 and to Pikes Peak in Colorado in 1911. The results from the later study suggested a 30% increase in hemoglobin mass, which however likely was a measurement artifact associated with the notorious difficulties associated with quantifying CO in blood at that time (they did not have automated blood gas analyzers back then…).
In the late 1940s brilliant but somewhat forgotten Swedish scientists Kjellberg, Rudhe and Sjöstrand determined blood volumes in humans using the CO re-breathing method and related this to cardiac output and exercise capacity. This was decades ahead of the later so famous “Scandinavian exercise physiology school”.
Some 50 years later the method received a facelift by Australian researchers Burke and Skinner in that they established a somewhat handier set-up. A summary and detailed technical description of the method can be found here.