INSTITUTE OF PHYSICS PUBLISHING EUROPEAN JOURNAL OF PHYSICS Eur. J. Phys. 26 (2005) 189–193 doi:10.1088/0143-0807/26/1/019 The sky is falling: Newton’s droplets, Clausius’s bubbles and Tyndall’s ‘sky matter’ Peter Pesic St John’s College, 1160 Camino de la Cruz Blanca, Santa Fe, NM 87505-4599, USA Received 20 July 2004, in final form 10 September 2004 Published 29 December 2004 Online at stacks.iop.org/EJP/26/189 Abstract To explain why the sky is blue, Newton, Clausius, Tyndall and Lodge invoked small suspended particles. However, simple calculations show that, though such particles fall very slowly, they are absorbed by larger ambient water droplets faster than they can be replenished. These estimates could have been made long ago to show that molecular scattering is necessary to explain the blue sky. 1. Introduction In the history of explanations why the sky is blue, several important theories involved suspended particles [1]. In 1704, Isaac Newton suggested that light passing through water droplets would give a ‘blue of the first order’ he identified with the ‘excellent blue of a bright sky-colour’ [2]. In 1849, Rudolf Clausius suggested that these droplets were really bubbles 1 [3]. But in either case, Ernst Wilhelm von Br¨ ucke noted in 1853 that the resultant blue would not be as deep (saturated) as sky blue [4]. Perhaps unaware of Br¨ ucke’s work, in 1854 Michael Faraday still advocated Clausius’s ‘vesicles’ [5]. In the 1860s, John Tyndall speculated that what he called ‘sky matter’, particles bigger than molecules but still submicroscopic, could explain sky blue and also the blue clouds he was able to make in test tubes by shining intense light into the fumes of volatile liquids such as butyl nitrate and benzol [6]. He speculated that sky matter might be water vapour or organic ‘germs’ [7]. In 1871 Lord Rayleigh suggested common salt crystals and in 1878 Oliver Lodge suggested dust from meteorites [8, 9]. In this paper, I will give simple order-of-magnitude estimates to show that such particles could not remain suspended, leaving molecular scattering as the only possibility. 1 The work of Clausius (1849) was discussed by Garber (1975) (second reference of [3]), which gives a complete list of Clausius’s publications on this question and a very helpful overview. For a treatment of his pioneering use of probability techniques in these studies see the third reference of [3]. 0143-0807/05/010189+05$30.00 c  2005 IOP Publishing Ltd Printed in the UK 189