• 1  Research presented in this paper was carried out under the post-doctoral grant SFRH/BPD/73373/2010 (...)

1Observatories played an important part in the development of European colonial expansion (Aubin et al., 2010, p. 2). Besides assisting navigation, metropolitan observatories supported colonial surveys, promoted the mapping of previously uncharted skies, and set the guidelines for the study of colonial climates. Colonial observatories were often founded as a consequence of these undertakings. In other cases, they emerged together with aspirations and agendas that equated local autonomy with scientific progress. All of these settlements were involved in complex networks that shaped the geographies of empire, as they nurtured the circulation of practitioners, instruments, ideas and techniques over vast geographical areas.

• 2 This expression is drawn from (Clarence-Smith, 1985). For a recent approach to the Portuguese empir (...)

2The establishment and development of a network of colonial observatories were an endeavour recurrently embraced by stalwarts of the Third Portuguese Empire2, which lasted from the 1820s (after the independence of Brazil) to 1975. This paper stems from preliminary research on this subject; it is intended to provide a tentative overview of Portuguese colonial observatories and their relations with the metropolis.

• 3 The Estado Novo (“New State”) was preceded by a military dictatorship that lasted from 1926 to 1933 (...)

3The provisional picture conveyed in the pages that follow develops around five key episodes: the attempt to create an imperial network of meteorological observatories and outstations controlled by the Infante D. Luis Meteorological Observatory (Lisbon) in 1857, the establishment of the Luanda Observatory (later João Capelo Observatory) in Angola in 1879, the inauguration of the Campos Rodrigues Observatory in Lourenço Marques (nowadays Maputo) in Mozambique in 1908, the attempt to upgrade the João Capelo Observatory in the 1920s, and the constitution of the National Meteorological Service of Portugal (based in Lisbon) in 1946. Approaching these undertakings requires us to go through distinct periods of Portuguese history: the liberal monarchy (1820-1910), especially the period known as Regeneração (1851-1890), the First Republic (1910-1926), and the Estado Novo (1933-1974).3 Imperial and colonial agendas changed in the course of these political regimes but, one way or another, they were always grounded in the mythical idea of reviving the glories of the First Empire – that is, the maritime and commercial empire founded in the sixteenth century – and thus compensate for the loss of Brazil, the great colonial domain of the Second Empire. Africa, and especially Angola and Mozambique, constituted the focal point of these ambitions. The observatories and observing networks of these former Portuguese colonies will receive especial attention.

• 4 The classic study by Bruno Latour (1987, ch. 6), Science in Action, and especially the concept of “ (...)

4Colonial observatories constitute the subject of a growing body of literature. They have been approached, for instance, as focal points for the introduction and development of instrumental sciences in the colonial context, as undertakings entangled in conflicting views on how to assemble a proper observatory, and as nodes of the logistic circuits of empire (Zuidervaart, 2004; Schaffer, 2012; Mcaleer, 2013). These issues can also be identified in what follows, but this paper is mainly concerned with a more general problem: the problem of action from a distance,4 that is, the attempts to steer an imperial network of observation from Lisbon, the tensions and obstacles that emerged in this undertaking, and the strategies implemented in order to overcome its predicaments.

• 5 Escola Politécnica de Lisboa.

5Between 1853 and 1854, Guilherme Dias Pegado (1804-1885), a physics lecturer at the Polytechnic School of Lisbon,5 steered the installation of a meteorological observatory in the premises of that school. Pegado was aware of the growing concerns with the standardisation of meteorological practices, an issue that was thoroughly discussed (albeit with limited success) at the international meteorological conference convened by Matthew Fontaine Maury (1806-1873) and held in Brussels in 1853 (Anderson, 2010, p. 245).

• 6 Observatório Meteorológico do Infante D. Luiz.

7In 1856, the observatory founded by Pegado was named Infante D. Luis Observatory6 (henceforth IDLO) after its patron Prince Luis (later King Luis I). The IDLO was meant to function as a centre of calculation for meteorology in the mainland. It was also assigned the supervision of meteorological observations carried out aboard Portuguese vessels (Simões et al., 2013, p. 106-111). There was still a lot to do at the metropolis; a meteorological network would have to be built from scratch. But the overseas empire was not forgotten. In April 1857, the Minister of the Navy, Viscount Sá da Bandeira (1795-1886) ordered the installation of meteorological stations in Cape Verde, Angola, Portuguese India, and Macau (a station had already been installed in Lourenço Marques). Sá da Bandeira had been campaigning for a “New Brazil” in Africa since the 1830s. As a keen purveyor of Portugal’s imperial rebirth, he did not hesitate to endorse this grand meteorological plan. Guilherme Pegado was required to supervise the assemblage of colonial stations from the metropolis. A decree dated 17 August 1857, also signed by Sá da Bandeira, determined that colonial governors should carry out a periodical inspection of instruments and observations, and present, at least, one report per year. Official documents from 1858 and 1859 mention the dispatch of instruments, as well as observing protocols and forms to Cape Verde, Bissau, Angola, Mozambique, and Goa (Ferreira, 1962, p. 6).

• 7 In 1871 the School was replaced with the Professional Institute of Goa, to which the observatory re (...)
• 8 Boletim Oficial do Estado da Índia, 2 October 1860.
• 9 Sociedade de Geografia de Lisboa.

9The meteorological activity might have developed more steadily in Goa, where an observatory was installed in 1860 at the Mathematical and Military School of the Portuguese Engineering Corps in India.7 In Macau, observations were made from 1862 onwards, first at the Military Hospital of the Peninsula, and later at the captaincy of Macau’s port.8 However, in the face of the “Scramble for Africa”, Portuguese authorities increasingly directed their attention towards the African colonies, and especially to Angola and Mozambique. In 1875, the Lisbon Geographical Society9 (henceforth LGS) was launched to promote the exploration of the African hinterland, which hitherto remained largely unknown (Medeiros, 2004). Portugal claimed to be the historical sovereign of those territories, but after the colonial conference held in Berlin in 1884-5, effective occupation became the sole criterion of sovereignty accepted by the other colonial powers.

• 10 Observatório João Capelo.

10It was in this context that an observatory was, for the first time, established in Portuguese Africa with a permanent character, within the institutional framework of colonial administration. The Luanda Meteorological Observatory was founded in April 1879, under the aegis of the Department of Public Works of Angola. Its regulations, issued in July 1879 (Government of Angola, 1879), prescribed a staff of two elements (the director and an assistant observer), who were expected to work in the observatory and to develop a network of meteorological stations. Each station was to be equipped with a barometer, thermometers, an udometer (rain gauge), an evaporimeter, a weather vane, an ozonometer, a print showing the different types of clouds, and some accessory devices. A detailed set of instructions for each kind of observation was also included in the regulations. All of these dispositions were sanctioned by the metropolitan authority of the IDLO, which, by that time, was directed by João Carlos de Brito Capelo (1831-1901). Educated as a Naval officer, Capelo becomes an observer at the IDLO in 1855. In 1875, he was appointed Director. Capelo’s works on ocean streams, terrestrial magnetism, and solar activity garnered him recognition in Portugal and abroad (Bonifácio et al., 2007). In 1916, the Luanda Observatory was rechristened João Capelo Observatory10 (henceforth JCO) in his honour.

• 11 For a comprehensive picture of the First Portuguese Republic and its ideological foundations, see ( (...)

13The Africanist agenda gained momentum in Portugal throughout the 1880s, but was seriously challenged in 1890. The Portuguese had been nourishing dreams of a great colony extending from the coasts of Angola to the shores in Mozambique, on the opposite side of the continent. However, England also intended to reinforce its African domain with a long railway connecting the Cape with Cairo, and did not want the Portuguese on the way. On 11 January 1890, Lisbon was compelled to order a retreat of Portuguese military forces deployed in disputed territories between Angola and Mozambique. The Portuguese authorities could do little more than to comply and seek a satisfactory deal in the ensuing negotiations (Teixeira, 1990). The episode, known in Portugal as the “British Ultimatum”, triggered perceptions of imperial fiasco and national abashment, putting the Portuguese liberal monarchy in jeopardy. But it also gave a new thrust to imperial aspirations in the country. Over the next two decades the Republican ideology would get stronger and stronger, until taking over in the revolution of October 5th, 1910. And so did the idea of imperial re-enactment, which was keenly incorporated in the Republican propaganda and increasingly perceived as a formula for national rebirth.11

• 12 For a biographical synopsis see (Canas, 2009, p. 69-79).

14The Portuguese techno-scientific elite was generally keen to embrace this ideal and to shape it into a programme of rational colonialism. Ernesto de Vasconcelos (1852-1930)12 featured among those who took on this task. A naval officer and hydrographical engineer by training, Vasconcelos paved a sound career in the politics and the administration of the empire. A founding member of the Lisbon Geographical Society, he became its Perpetual Secretary in 1911. This appointment allowed him to consolidate a persona he had been crafting for years: that of a mastermind commanding the techno-scientific resurgence of the Portuguese empire.

• 13 (Castelo Branco, 1906). On the relations between Mozambique and South Africa in this period see (Ka (...)

17 Lourenço Marques’ port was crucial for Mozambique’s economy. Together with a railway connecting the Mozambican capital with Johannesburg, it constituted a major gateway for people and goods in South-eastern Africa. After the South-African War of 1899-1902, the ports of the Cape and Durban emerged as serious competitors. The programme commanded by Lacerda was meant to secure the competitiveness of the Portuguese port.13 Lacerda also saw it as an opportunity to present Portugal as a sophisticated colonial overlord, for which the establishment of an observatory in the port seemed very appropriate.

• 14  Observatório Astronómico de Lisboa.

18The Luanda Observatory had been officially in activity for a quarter of a century now, but it was not in Angola that Lacerda sought inspiration for his endeavour. Similarly to Vasconcelos, he had practised at the IDLO and at the Astronomical Observatory of Lisbon14 (henceforth AOL) as part of his training as a hydrographical engineer. Lacerda maintained close ties with the AOL thereafter.

• 15  On the foundation and history of the AOL until 1910 see (Raposo, 2010).
• 16 This time-ball was installed in 1885, in replacement of an older (and very inaccurate) time-ball th (...)

19The AOL was founded in 1857 to collaborate with the Pulkovo Observatory (near St. Petersburg, Russia) in the development of stellar astronomy, namely in the measurement of stellar parallax. However, due to delays in its construction and to a shortage of qualified personnel, it was eventually refashioned as an observatory devoted to timekeeping, and accessorily to the measurement of stellar positions.15 From the 1880s onwards the AOL firmed its status as a national timekeeper, transmitting the official time to a time-ball in the port of Lisbon.16 The exactness of the Lisbon time signals became a badge of the AOL’s commitment to precision. The astronomer César Augusto de Campos Rodrigues (1836-1919), also a Naval officer and hydrographical engineer by training, gave an important contribution to this state of affairs, with his technical contraptions and studies on observing methods. In 1904, he was awarded the Valz Prize of the Academy of Sciences of Paris. The prize consolidated his image as a local scientific hero, and firmed the AOL’s status as an internationally respected centre of positional astronomy. This imbued Lacerda with the confidence necessary to engage in a transfer of timekeeping technology from the AOL to the remote harbour area of Lourenço Marques.

• 17 Serviço Meteorológico dos Açores.

20 The AOL was in its heyday, contrasting with the IDLO. Over the previous decades, the IDLO had functioned as a one-man-observatory. It fell into decline after the demise of its virtuoso meteorologist João Capelo in 1902. The focal point of Portuguese meteorology gradually diverted towards the Azores, where the Army officer and science enthusiast Francisco Afonso Chaves (1857-1926) had recently established the Meteorological Service of the Azores17 (henceforth MSA), based at the Ponta Delgada Meteorological Observatory (Tavares, 2009). Chaves took advantage of the importance of Atlantic data for weather forecasts in order to firm the Azores as a major node in the international meteorological networks, and to fashion himself as a spokesman for Portuguese meteorology, in which he succeeded.

• 18 “Observatório Campos Rodrigues”.
• 19 “(…) natural incentivo ao derramamento da instrução” (Observatório Campos Rodrigues, 1909, p. 7).

23In 1908 the new observatory was inaugurated in Lourenço Marques and officially named Campos Rodrigues Observatory.18 It was located near to the port, in an area of public estate, beside a major access to the city (Polana road). Lacerda conceived the observatory as a local monument of science: he wanted the surrounding parks to be freely accessed by passers-by, so that they could experience “a natural incitement to a burst of instruction”.19

• 20 A very schematic and succinct journal written by Oom during this mission is kept at the Historical (...)

25After some initial setbacks, time signals started to be sent regularly, via the telegraphic network, to a clock installed in the boarding area of the port. The signals were also relayed to a system of luminous semaphores that displayed the time in the freight area. This arrangement replicated the system of public time signals that Oom had saw at the Hamburg port. It was eventually adopted in Lisbon as well.20

• 21 Freire de Andrade held the post of Governor of Mozambique between 1906 and 1910. The observatory wa (...)

27The CRO also contributed with data for the study of latitude variations promoted by the International Geodesic Association (which Portugal had joined in 1867). However, the astronomical section remained essentially focused on timekeeping. Its apparatus did not allow for much more. Not all Portuguese dignitaries shared Lacerda’s enthusiasm for the observatory. Mozambique’s Governor Freire de Andrade (1859-1929), a military engineer and stalwart of colonial development,21 questioned the CRO’s relevance for the local economy. As a consequence, the colonial authorities provided limited funding. Lacerda and Oom were forced to order a Bamberg transit instrument, much less reliable than the Repsold instruments used for decades at the AOL. They also had to buy cheaper clocks, a predicament that Campos Rodrigues and Frederico Oom sought to circumvent by examining, testing, and whenever possible, improving the equipment before shipping it to the CRO.

• 22  In Lisbon, Krille and Max-Richter clocks were used to keep, respectively, sidereal time and mean s (...)
• 23  The acquisition of the CRO’s instruments is documented in the above-mentioned file C648 of the His (...)

28However, the behaviour of instruments in the remote lands of South-eastern Africa was often deviant from their performance in the metropolis. This is well illustrated by the first mean solar time clock of the CRO, ordered from the Munich-based firm Riefler. Campos Rodrigues and Oom knew in advance that it would never perform at the same level as the AOL’s clocks,22 but tried to improve it as much as possible nevertheless. After thorough tests at the AOL they deemed it able to provide a satisfactory level of precision. But once installed at the CRO, its march proved so irregular that there was no option other than acquiring another clock.23 In these circumstances, replacing or fixing an instrument usually required sending it back to the metropolis, from where it would be redirected to an expert firm abroad, or from where an order for a replacement would be placed and processed. The whole circuit of supply, test and delivery, which involved Lisbon, instrument workshops located elsewhere in Europe, and the Portuguese colonies, had to be re-enacted. The picture was worsened by the absence of advanced instrument and clockmakers in the Portuguese mainland itself.

• 24 The Faculty of Sciences of Lisbon was established in 1911 in the sequence of the Republican Revolut (...)

30It was only from 1916 onwards that the CRO’s staff found some stability, with the appointment of Manuel Peres Júnior (1888-1968) to the post of director. Peres had practised for two years at the AOL, after obtaining a degree in mathematics from the Faculty of Sciences of Lisbon.24 Better prepared and more dedicated than his predecessors, Peres consolidated a tradition of accurate timekeeping in the colony, whilst promoting the development of the Mozambican meteorological network.

• 25 On isobars and meteorological mapping see (Anderson, 2010, ch. 5).
• 26 See the reports of the CRO (Relatório do Observatório Campos Rodrigues em Lourenço Marques) for the (...)

31The meteorological equipment of the CRO comprised five basic sets of instruments, for the following measurements: air pressure; measurements in the shade (temperature, evaporation and humidity); precipitation and soil temperature; direction and intensity of winds; actinometry (study of solar irradiation). The apparatus also included some self-recording instruments, namely a thermograph, two barographs, a psychrometer (for humidity), an udograph (for precipitation), and a Jordan heliograph (an instrument that recorded sunlight on an especial type of paper) (Observatório Campos Rodrigues, 1909). A forecast system based on isobars was eventually implemented.25 The data used in the production of the forecasts was collected through a network of meteorological stations that started to expand after the Great War. It counted on stations at port captaincies, administrative services, private companies, and also on amateur contributors. Peres placed a considerable effort in disciplining station observers, distributing standardised instructions and forms, offering apprenticeships at the CRO, and carrying out inspections in situ whenever possible. However, it was only with the professionalization of observers and the introduction of financial prizes in the early 1920s that the observing routines in the network reached a satisfactory level of stability and reliability.26

• 27 Circular No.15, 1914, January 19, of the Union Observatory; Augusto Teixeira, “Diferença de longitu (...)

33The news certainly pleased Lacerda, who, by this time, was in Macau commanding a new harbour venture. Garnering recognition from abroad was the utmost achievement he had desired for the Mozambican observatory. In fact, Lacerda had conceived the CRO as a catalyser of scientific liaisons between Mozambique and South Africa. As interim director of the CRO, a post he held between 1908 and 1909, he obtained the agreement of the Johannesburg Observatory to standardise meteorological telegrams in the whole region (Observatório Campos Rodrigues, 1909, p. 9); he also promoted joint longitude measurements and barometric investigations.27 With the aim of strengthening these ties, Lacerda suggested that the South African Association for the Advancement of Science (SAAAS) hold one of its annual meetings in the Mozambican capital. The SAAAS had been founded in 1903, reflecting the importance of science and technology to the ideology of South African autonomy (Dubow, 2006)).

• 28 The programme and the proceedings of the conference are presented in (SAAAS, 1913).

34Lacerda’s idea was well accepted by the Association. Between the 1st and the 13th of July 1913, Lourenço Marques hosted the eleventh meeting of the SAAAS. The meeting evinced a blatant imbalance between the scientific milieus of the two colonial domains. While South-Africa sported various observatories, museums, and other cultural institutions where the sciences were enthusiastically cultivated, Mozambique had little more to display than the CRO. This imbalance was evident in the proportion of papers presented by South-African and Portuguese (or Mozambique-based) attendants, which was clearly favourable to South-Africa.28

• 29 “Quasi nada ha feito, até agora, de verdadeiramente scientifico e manifestamente util, com respeito (...)

37Another critic of the CRO was Manuel de Brito Camacho (1862-1934), a prominent Republican who governed Mozambique between 1921 and 1923. In 1926 Brito Camacho wrote that “almost nothing really scientific and clearly useful has been done with respect to meteorology in the colony, albeit there is in Lourenço Marques an observatory entrusted with these investigations.”29 It was true that by then the CRO was still dealing with a shortage of station observers, and striving to extend its coverage of the hinterland. Still, Brito Camacho’s remarks were unfair. Given the circumstances, the CRO had reached a noteworthy level of performance, especially when compared to the older observatory of Angola.

• 30 Norton de Matos did not run in the end, arguing that the elections would be manipulated by the regi (...)
• 31 For an overview of Norton de Matos’s political action in Angola see (Wheeler, 2009, p. 171-175).

38In 1912, José Ribeiro Norton de Matos (1887-1955) was appointed governor of Angola. Norton de Matos is best remembered in Portugal for daring to present himself as a democratic presidential candidate against Oliveira Salazar’s dictatorship, the Estado Novo, in 1948.30 This initiative granted him an enduring aura of freedom’s paladin; his colonial legacy is more controversial. The efforts he placed at modernising Angola are undisputed. But his action as colonial governor is also marred by a seeming penchant for authoritarianism, coupled with an ambivalent attitude towards native populations. Norton de Matos showed concern for their working conditions and stood for their inclusion in the colonial administrative apparatus. However, he did so always from a strongly Eurocentric perspective, grounded on equal amounts of racism and paternalism. The education and assimilation of natives should, in his view, be carefully limited so that they did were not seduced by independentism. Whenever they were and showed it, Norton de Matos responded with stark repression.31

• 32 “Restabelecer em tudo que nos diz respeito a inteira verdade histórica: - que nos baste sempre a me (...)

39Norton de Matos’s faith in the imperial destiny of Portugal was apparently fierce, and so was his trust in science as a tool to fulfil such a destiny. A military surveyor by training, he firmly believed, like Vasconcelos, Lacerda and other contemporaries, that the erstwhile glories of the Portuguese maritime empire could be revived through a steady programme of rational colonialism focused on Africa, which he often referred to as “scientific occupation”. To promote science in the African colonies was, in Norton de Matos’s own words, to “re-establish, in everything that concerns us [the Portuguese], the historical truth”32 that is, the continuation of a civilising mission deeply inscribed in Portugal’s fate as an imperial nation.

• 33  For a detailed analysis of Norton de Matos’s first tenure in Angola, see (Dáskalos, 2008).
• 34 “Circular do Governador Geral de Angola, de 17 de Abril de 1913”, in (Matos, 1953, p. 155-156).

40Norton de Matos first governed Angola between 1912 and 1915. He returned in 1921 as High-Commissioner of the Portuguese Republic (a post equivalent to governor), in which he stayed until 1924.33 During his first tenure, Norton de Matos sought to reverse the generally backward state of the colony by reorganising its administrative framework: the departments for public works and agriculture, the customs, the mail and telegraph offices, among other services (Government of Angola, 1921, p. 55). His techno-scientific concerns dwelt primarily in the colonial health system and the surveying department (Government of Angola, 1921, p. 54), but meteorology was not forgotten. In a circular dated from April 1913, Norton de Matos demanded the higher civil servants of Angola to maintain a decent and “European” appearance in the premises they oversaw, to be aware of all legislation related to their assignments, and to obtain all instruments necessary to characterise the geography of their circumscriptions, which included their climatology.34

• 35 Unidentified sender to Campos Rodrigues, 21 January 1911, Historical Archive of the AOL, file DD601

42As far as astronomy was concerned, the time service of the JCO was seemingly in a deplorable state. In 1911, the AOL was informed that a Bamberg transit instrument used to carry out time observations had reached an advanced state of disrepair, and that the time-ball system was “the most primitive one”: an assistant pulled a rope to release the ball upon hearing a ring clumsily activated by a clock. This arrangement rendered an average error of three seconds in the time signal,35 which for the AOL astronomers was outrageous. The AOL published a quarterly report with the errors of the Lisbon time-ball to cents of a second. Frederico Oom liked to flaunt that the error of the Lisbon time signal rarely went above a few tenths of a second – a widespread benchmark of chronometric acumen in the late nineteenth century and the early twentieth century (Canales, 2011), which the AOL had militantly embraced.

• 36  Frederico Oom, “Novo Observatório em Luanda”, 30 September 1922, Historical Archive of the AOL, fi (...)

44Colonial medicine was the focal point for techno-scientific liaisons, which left the JCO in a secondary position. But Norton de Matos also had ideas for the observatory. In a speech delivered on 15 March 1922 (Matos, 1926, p. 164), he declared that the JCO and its meteorological services required an urgent upgrade. A new supply of instruments was eventually demanded from the metropolis, in order to re-enact dismantled stations and develop the network. The High Commissioner also wanted the astronomical section to be revamped, for which he summoned Frederico Oom to Angola. The astronomer visited the JCO in August 1922. In his report on the visit,36 Oom stated it was pointless to compete with well-established observatories such as those of Durban and the Cape; the Portuguese authorities should focus instead on a reliable time service instead, using the CRO as a model. A second of precision was, according to Oom, the minimum goal to be attained, if Luanda wanted to sport a trustworthy time signal. Portugal’s prestige was at stake, he further remarked, as Luanda’s time service was the only one available in the region. Ideally, an effort should be made to work on tenths of seconds. Consequently, Oom added, the JCO should be transferred to purpose-built premises. And time observations with a theodolite, a makeshift solution that had been adopted in place of the deranged Bamberg transit, were to be discarded altogether.

• 37 These instruments included the already mentioned Bamberg transit, a 108-mm Bardou telescope, 3 chro (...)
• 38 The problems surrounding the acquisition of these instruments are partially documented in the file (...)

45In the sequence of Oom’s visit, several astronomical instruments were sent from the JCO to Lisbon for repair. Most of them were deemed useless, or, at best, to contain some reusable components.37 A set of new instruments was ordered in 1923. It included a new Bamberg transit, a Max-Richter clock, and a Nardin chronometer. Once again, the instrumental circuits of the empire proved difficult to deal with. The chronometer was successfully tested in Lisbon, but in Luanda its march revealed irregularities. A clockmaker based in the Angolan capital was unable to repair it, thus the chronometer had to be sent back to Lisbon. A local clockmaker managed to fix it, but some accessories had to be ordered from the Nardin firm in Switzerland. The chronometer returned to Luanda in 1928 only. There was also trouble with the new Bamberg transit. Oom had asked Bernhard Wanach (1867-1928) of the Potsdam Geodetic Institute to test the instrument there, which caused a delay in its dispatch to Lisbon. And when the instrument was finally unpacked in Luanda in 1925, the personnel of the JCO were desolate to find that some components were missing. In the case of the clock, the situation was even more dramatic, as both the clockmaker and Wanach died before it was dispatched to Angola.38

• 39 Vasco Lopes Alves to Frederico Oom, 23 April 1925, Historical Archive of the AOL, file DD601.
• 40 J. A. Bacelar[?] to Frederico Oom, 16 September 1913, Historical Archive of the AOL, file DD601.
• 41 J. A. Bacelar[?] to Frederico Oom, 18 January 1923, Historical Archive of the AOL, file DD601.

46Further research is needed to clarify what happened in the ensuing years, but it is certain that the assemblage of the new observatory was yet to be completed when Norton de Matos ended his second appointment in 1924. The construction of new premises was postponed due to financial difficulties39 and workforce issues. European workers were scarce and generally regarded as cunning opportunists, whilst natives were deemed lazy and untrustworthy.40 It was not easy to recruit personnel for the observatory either, a situation that changed only when wages were raised.41

• 42  Respectively: Anuário Meteorológico do Observatório João Capelo; Observações meteorológicas de alt (...)

47Significant improvements seem to have taken effect only by the late 1930s. In 1937 a new publication was launched: the Meteorological and Climatological Elements of Angola (Elementos Meteorológicos e Climatológicos de Angola). Angola’s meteorological network grew substantially in the ensuing years. By the early fifties, the JCO controlled more than 180 stations (Serviço Meteorológico de Angola, 1954, p. 7-8; Ferreira, 1952, p. 15) and presented its results in three different publications: annual proceedings, tables of high altitude observations with balloons, and reports of ground observations.42 A time service with radiotelegraphic transmission of time signals was also at work.

• 43 For a comprehensive picture of the Estado Novo, see (Mattoso & Rosas, 1998).

50In 1933, the Constitution of the Estado Novo was formally approved, firming the ground for a dictatorial regime that would last until 1974.43 Fiercely commanded by António de Oliveira Salazar (1889-1970), the regime used the political instability of the First Republic (between 1910 and 1926 there were 39 cabinets and numerous episodes of political violence) as a major argument to impose a strict order, grounded on political surveillance and repression. Empire was embraced as a badge of grandeur and might, and reviving its erstwhile glories became one of the Estado Novo’s foundational myths (Rosas, 2012, p. 34).

• 44 It must be remarked that the parliament of the Estado Novo (Assembleia Nacional) was essentially an (...)

54Once appointed director of the IDLO, Ferreira immediately sought to activate a legal disposition issued in 1923, which established a National Climatological Service under the guidance of the IDLO. This provided him with a leverage point to extend his influence and authority over all other meteorological services in Portugal. The fact that Ferreira was very well placed in Salazar’s regime was obviously of great advantage. He was a member of parliament44 for many years, and the Sub-Secretary of State for Education between 1944 and 1946. It was around this time that Salazar’s government entrusted Ferreira with the reorganisation of Portuguese meteorology, an issue that, according to a biographer, the physicist discussed thoroughly with the dictator himself (Peixoto, 1980, p. 67).

• 45 Serviço Meteorológico Nacional.

55The result was the National Meteorological Service45 (SMN), founded in August 1946. Amorim Ferreira was promptly appointed Director, a post he maintained until retiring in 1965. Henceforth all meteorological work in the Portuguese mainland, the islands of Madeira and the Azores, and the overseas colonies was to be closely supervised by the SMN. In 1950 all colonies were officially endowed with a meteorological service constituting a sub-department of the SMN (Ferreira, 1952, p. 14). In the metropolis, all meteorological and geophysical observatories (IDLO, eventually renamed as D. Luis Institute of Geophysics; the Institute of Geophysics of the University of Porto, former Queen Amélia Observatory, and the Geophysical Institute of Coimbra) were equally placed under its aegis (Ferreira, 1952, p. 10). The constitution of this new meteorological overlord was accompanied by an effort to redefine the professional training and status of meteorologists and other practitioners of the Earth sciences. For that purpose, degrees in geoscience were established at the universities of Lisbon, Coimbra and Porto. Students enrolled in these new programmes were expected to take posts at the colonial observatories upon obtaining their diplomas (Ferreira, 1962, p. 29).

• 46 On Ernesto Gherzi see (Udías, 2003, p. 123-124; Bell, 1974).

56Angola and Mozambique remained the centres of the network, but the activity of other observatories in the Portuguese empire seems to have gained momentum as well. By the early 1950s, a central observatory in the Island of Sal, Cape Verde, controlled an ancillary observatory installed in Mindelo (S. Vicente Island), 30 stations for ground observations, and an outpost for aerial observations at Praia (Island of Santiago). The Cape Verde Observatory also supervised an ancillary observatory established in Bissau (Portuguese Guinea), which served as the main node for a twelve-station network. In the far-eastern edges of the empire, the Goa Observatory commanded 15 stations (Ferreira, 1952, p. 15-16). The Macau Observatory got an additional impulse in the 1950s from the Italian meteorologist Ernesto Gherzi, S. J. (1886-1973). Gherzi was a keen investigator of storms in the seas of China, and the former director of the Zikawei Observatory in Shangai. He stayed in Macau between 1950 and 1954, before moving to Canada.46 The Portuguese government hired him during that period to steer the reorganisation of Macau Observatory, where, besides implementing the dispositions established by the SNM, Ghenzi also conducted investigations on cosmic rays. By the early 1960s, the Macau Observatory controlled a five-station network, deemed adequate to its small territory (Ferreira, 1960, p. 10). In East Timor, meteorology had been brought into a halt during World War II, as a consequence of the Japanese occupation of the island. All equipment and records maintained by the Portuguese authorities were lost at that time (Ferreira, 1960, p. 29). By the mid-1960s, Timor’s meteorological network had been fully re-established and significantly extended. It comprised 26 stations and was controlled by an observatory located in the capital, Dili (Ferreira, 1960, p. 29-30). All of these colonial services included at least one station for aerial observations and maintained forecasting services.

• 47 Marcelo Caetano succeeded Salazar in 1968 as President of the Governmental Council, which in practi (...)
• 48  Due to its special statute, Macau returned to Chinese administration in 1999 only. In East Timor t (...)

57In the metropolis, the SNM kept on amassing and publishing the data submitted by the colonies, as the IDLO and the Ministry of the Colonies had done prior to its foundation. Hitherto the function of the metropolitan “centres of calculation” had been essential to accumulate data from all over the empire. Apart from the elaboration of summary tables, few “calculations” had actually been performed. The development of the colonial networks and their tighter control from the metropolis put the SNM in a good stead to change this picture. In 1965, Amorim Ferreira presented a comprehensive study of dynamical climatology concerning Southern Africa (Ferreira, 1965), with data collected through the Mozambican and Angolan networks. Originally requested by the Portuguese Air Force, this work was crafted by Amorim Ferreira to serve a wide range of purposes: aerial navigation, agriculture, tourism, public health and urban planning, among others. The old dream of a steadfast imperial centre of meteorology in Lisbon, planning and fostering colonial development on the grounds of comprehensive climatological studies, was seemingly coming true. But as imperial meteorology started to get sounder, the empire began to crumble down. Salazar and his successor Marcelo Caetano (1906-1980)47 fiercely resisted the post-war wave of decolonisation, to no avail. In 1961, Portugal lost its territories in India (Goa, Damão e Dio) after their occupation by Indian troops. In the same year, the Estado Novo launched a war against pro-independence groups in Guinea-Bissau, Angola and Mozambique. The war went on until the “Carnation Revolution” of April 25th, 1974, which finally put an end to the Estado Novo. This was also the end of the overseas empire. It was officially dissolved in 1975.48 The former colonial observatories and their networks were eventually transformed into the national meteorological services of the new independent nations that emerged from the Portuguese Empire.

• 49 On the observatory of Mauritius (Royal Alfred Observatory) see (Kirk-Greene, 2001, p. 170-171; Macm (...)
• 50 On the history of these observatories see (Mcaleer, 2013; Evans, 1988, p. 107-109; Warner, 1979).

62It must be noted that the CRO and the JCO were not the only African colonial observatories where meteorology took over astronomy and other observatory sciences. The observatories of Mauritius and Madagascar, for instance, followed a similar path.49 But in the neighbour South-Africa astronomy got a solid footing, especially at the observatories of the Cape and Johannesburg.50 In fact, the latter began to function mainly as a meteorological observatory and ended up as a respectable centre of astronomy. Various factors contributed to this state of affairs: the solid traditions of astronomy and instrument making in the centre of the empire, Britain; the relatively early foundation (1820) and steady development of the Royal Observatory of the Cape; a network of learned institutions fomenting a south-Africanist ideology grounded in science; and the prestige garnered by astronomers such as David Gill (1843-1914) at the Cape and Robert Innes (1861-1933) in Johannesburg. The contrast with the Portuguese colonies is paramount.