Introduction: On November 17, 1993, Professor Hou Wenfeng, Director of the National Marine Data Information Service of the State Oceanic Administration of China and Gregory W. Withee , U.S. Chairman for the National Oceanic and Atmospheric Administration (NOAA) signed a Joint Implementation Plan to digitize the Maury Collection of historical ships' weather logs stored in the National Archives and Records Administration (NARA, 1981) in Washington D.C. The implementation plan was established under the aegis of the United States People's Republic of China (US-PRC) Protocol on Cooperation in the Field of Marine and Fishery Science and Technology, Joint Coordination Panel for Data and Information Cooperation. This agreement resulted in approximately 1.4 million ship observations being digitized for the general period between 1820 and 1860. There were eighty-eight reels of microfilmed records in the collection. Reels One, Two, and Forty-four do not appear on the CD-ROM. Reel One is simply an inventory, and we were unable to produce clear paper copies from Reel Two. Reel Forty-four, keyed in a different format, was used in the pilot digitization project to help finalize the formats that were adopted for the remainder of the collection. Reel Forty-four will later be converted to a common format. The files on this CD-ROM will be converted to a common format for inclusion into the Comprehensive Ocean-Atmosphere Data Set (COADS) which will be comprised of over 100 million records. Once the Maury data are merged, COAD's period of record will span from the late 1700s to the near present, with plans to update it every few years. Placing this collection on CD-ROM in the original digitized format will preserve the data and assist in the development of specifications for converting the records to a COADS formatted version. This initial version of the data will be reviewed and modified before being finally merged into the main COADS database. The successful digitizing of the Maury Collection would probably not have been accomplished, at least in this century, without the insight and guidance of Joseph O. Fletcher, former Assistant Administrator of NOAA's Office of Oceanic and Atmospheric Research. Dr. Fletcher was a key supporter of the critical need for digitizing early historical data to fill gaps in existing digital marine records. Groundwork for this project in the U.S. was also prepared by Roy Jenne, Henry Diaz, Ralph Slutz, Robert Quayle, Scott Woodruff, and Joe Elms, among others. In China, the Maury Project work was directed by Guo Fengyi and Zhang Dongsheng at the National Oceanographic Data Center of China (CNODC). In the U.S., Joe Elms coordinated the project at the National Climatic Data Center (NCDC). Once the data were digitized, much effort went into the production of the CD-ROM at NCDC. Thanks to Debra Braun for processing the data and generating the CD-ROM master copy, and to Scott Miller and Ann Hallyburton for the cover and CD-ROM label designs. Thanks also to Robert Quayle and Michael Changery for reviewing and commenting on the text. Background: The Maury Collection of ship observations housed in NARA and the resulting Wind and Current Charts produced by Matthew F. Maury based on these early ship reports (e.g., Maury, 1854) would NOT exist if not for a stagecoach accident in 1839 that left Lt. Maury permanently disabled. Knowing that his broken right leg and dislocated right knee would likely keep him from sea duty, Maury still requested assignment to the Pacific Squadron. His request for sea duty was never granted, but on July 1, 1842, he was assigned as officer-in-charge of the Depot of Charts and Instruments. This was not a very prestigious post and one that most officers would have considered an end to their career. The major responsibilities of the Depot prior to Maury's posting was to determine solar and sidereal time for calibrating Navy chronometers. Maury found a number of old ship logs stored in the Depot containing information on winds, currents, and storm frequencies. He remembered the difficulty in finding such information about the winds and weather for a number of regions he had sailed to early in his career and quickly recognized the potential of these ship logs. He soon had his staff surveying the old ship logs stored at the Depot for critical information which eventually led to the publishing of the first chart for the North Atlantic in 1847. He persuaded seamen to provide additional data by attaching special log sheets to the charts he was providing free of charge and requesting that they provide their observations in return for free future charts. At the bottom of some of the distributed abstract logs he had the following message printed, "You are expected in conformity with the agreement as per the foregoing receipt to send to the Observatory the abstract of every voyage you may make until the charts are completed. Vessels that fail to return abstracts will forfeit their claims to the charts." This was very successful and by the early 1850's, the Observatory was receiving thousands of reports from the world's oceans because mariners had come to value his climatological charts so highly. The success of these charts was measured by the increased safety and speed of the ships utilizing Maury's charts. In one logbook the entry indicated how important the contract was to the Ship's Master in supplying the weather observations in exchange for Maury's charts. The Captain returning to Norfolk from a trip to the West Indies writes on September 7, 1852, "I am sorry to say there is not much information gained on such a voyage out of Norfolk as our vessels are not any or very few of them furnished with barometers or thermometers and very few with even chronometers." He goes on to say how every Ship Master ought not be without a chart, so he is sending in his weather information according to the signed contract. The log contained no instrumental information, only wind directions (first, middle, and latter parts of the day) and remarks as to the wind strength and weather conditions (e.g. fresh breeze and squally). There was also a position (latitude, longitude) reported with each day's entry. Many of the early logbooks in the Maury Collection did not contain instrumental readings and were typical of this example. A few only contained wind direction, occasional ship positions and brief remarks regarding the general weather. Interest by other maritime nations in Maury's success in the new science of physical geography of the sea, and sponsorship by the U.S. Navy, led to an international conference in Brussels in August of 1853. In attendance were representatives from ten major maritime nations. The Conference was quickly followed by an enlarged and improved Sailing Directions publication (Maury, 1854), which included ninety pages of new oceanographic information. Also included were the minutes from the meeting in Brussels, containing a newly devised uniform system of meteorological observations at sea, an important metadata source. War has proven detrimental to the archiving of marine observations during the periods of conflict, e.g. World War I, World War II and, for the Maury Collection, the Civil War in the United States. During this period, Lt. Maury, originally from Virginia, left his post and resigned his commission in the U.S. Navy to return to the south, even though he opposed slavery and the separation of the Union. He served in several positions in the Confederate Navy during the Civil War. When the Confederacy lost the war, he remained outside the United States until a general amnesty was issued in 1868 by President Johnson, at which time he returned to his beloved Virginia. Maury then taught and worked the academic lecture circuit until his death in February, 1873. There have been a number of biographies and articles written on Maury which can provide more detailed information on his life and work (C. Lewis, 1927; H.Hawthorne, 1943; F.Williams, 1963; Bates and Fuller, 1986; and J. Lewis, 1996). Maury had been labeled the "Pathfinder of the Seas" and many of the incoming logbooks were addressed to the Pathfinder of the Sea, even years after he had left the post. The work Maury began has been carried forward, but often with varying degrees of emphasis depending on scientific direction and national and international politics. After Maury's legacy the next major work in the U.S. came nearly three quarters of a century later with the development of the Atlas of Climatic Charts of the Oceans (U.S. Department of Agriculture, 1938). It is often referred to as McDonald's Atlas after the Principal Meteorologist in Charge of the Weather Bureau Office in New Orleans, Louisiana, Willard F. McDonald, under whose supervision the atlas was prepared. The work began in 1934 as a Civilian Conservation Corps/Work Projects Administration (CCC/WPA) project, comprising nearly 5.5 million observations in its database, mainly covering the period from 1885 through 1933, and at times employed 250 personnel. Many of these observations came from the original Maury logs. It appears that only three hundred and fifty-five of Maury's five hundred and thirty-three volumes were among the records transferred to the National Archives in 1938. Of those three hundred and fifty-five volumes, three hundred and forty-three had been included in the WPA project in 1936. The WPA project merged these three hundred and forty-three logbooks with one hundred and eighty logs not associated with Maury, and re-numbered the volumes thus losing the identity of the original Maury logs. The data came from many areas of the world's oceans, although confined mostly to the major trade routes, allowing the combined analysis to cover most of the globe with the exception of the polar regions. The observations came from many maritime nations, although the United States, United Kingdom, Netherlands, Germany, Japan and France were the major contributors. This new collection of five hundred and twenty-three logs has become known as the Maury Collection and is the material digitized under the project completed in 1996 between CNODC and NOAA. After World War II, the need for an updated global ocean atlas was quite evident. Therefore, during the 1950s the U.S. Navy sponsored the development of a series of ocean basin atlases covering the globe. These were updated later in the 1970s through sponsorship of the U.S. Navy, and were issued in 1992 as a Navy CD-ROM Global Atlas. Since then more emphasis has been placed on climate and global change studies. With an increased recognition for the importance of long-term marine records and their potential for climate research (e.g., Fletcher et al.,1981), inter-agency cooperation began in January 1981, when NOAA's Environmental Research Laboratories (ERL), and the Cooperative Institute for Research in Environmental Sciences (CIRES) planned with NCDC to combine additional marine data sources that had become available. The basic data set at the time was referred to as TDF-11 (NCDC, 1968). The National Center for Atmospheric Research (NCAR) soon joined in planning and execution of this continuing project, which produced the Comprehensive Ocean-Atmosphere Data Set. COADS (Slutz et al., 1985; Woodruff et al., 1987; Woodruff et al.; 1993; Woodruff et al., 1997) has continued to be updated through this cooperative arrangement, including vital international contributions such as the digitizing of the Maury Collection by CNODC, data exchanged under the auspices of the World Meteorological Organization, and other bilateral data exchanges. Quality controlled ship reports now covering the period from 1854 to 1995 are available in COADS, having been supplemented in more recent years by near-surface oceanographic sea temperatures and data from drifting and moored buoys. The Maury Collection (primarily 1820-1860), the 1912-1946 U.S. Merchant Marine Collection, and other international data sets will be converted to a common format, designated quality control steps will be completed, and the resulting data added to COADS. Maury Data Preparation and Characteristics: The Maury Collection was copied onto eighty-eight reels of microfilm (NARA, 1981) to facilitate preservation of the paper log books and for external distribution. NCDC received the complete series of microfilm reels from which it produced oversized paper copies. The copies were sent to CNODC in Tianjin where they were digitized. Nearly every salvageable record in the Collection was keyed. At the time the Maury Collection was microfilmed by NARA in the early 1980s, the filming equipment available could not automatically sense the proper setting required to obtain an optimal copy of each page as is possible with today's equipment. Thus the settings were checked and established only once at the beginning of each reel to be filmed, and remained constant throughout the approximate one thousand frames of filming. Therefore, when exceptionally weak log book pages were encountered, a very poor quality frame was produced. Some of these were of such poor quality that they could not be reproduced for forwarding on to China for digitizing and therefore, have been lost to the digital database. There were also some frames on the microfilm that did not deal with weather information (e.g. biological information) and these were not copied for shipment to China. Even some of the copies sent to China were impossible to read in certain sections resulting in the inability to digitize the information. It is estimated that less than five percent of the data were lost because of filming problems. It is clearly evident in the post-keying inventories, that Maury was able to collect an increased number of records during the 1850s, at the height of his career as Superintendent of the Depot of Charts and Instruments (which under his tenure became the U.S. Naval Observatory). The precipitous drop in the collection of observations with the onset of the U.S. Civil War is also evident. In addition to the variable reproduction quality, the penmanship was often very difficult to read and interpret, which also led to some loss of information. The mariners would often use abbreviations in entering their observations into the log books. Some of these were rather difficult to identify especially when the handwriting was difficult to read; examples include "sp wh" referring to a sperm whale, "sqs" or "sqls" for squalls, "mod" or "mdt" for moderate, "lt" for light , "st" for strong and "fr" for fresh breeze. Notations first interpreted as "2o clear" or "3o rain" initially appeared not to make sense. However, after considerable review of a number of different page entries, a few of the symbols that first appeared to be degree symbols seemed to more closely resemble a small lower-case "d". Since every example was preceded by a "2" or "3" we finally concluded that "2d" and "3d" indicated the second or middle part of the day and third or latter part of the day. This then fit the general practice for entries in the "Remarks Section" as there was generally a first entry followed by a "2d" and "3d" entry, or a first entry followed by one other entry ("2d" or "3d"), each including a short description of the weather. For example, for a given day the entry might appear as "clear, 2d rain, 3d fog"; this indicated clear conditions during the first part of the day (noon-2000), with rain occurring during the middle part of the day (2000-0400) and finally followed by foggy conditions during the latter part (0400-noon) of the day. Another record might appear as "rain squalls, 3d thick haze", indicating rain showers during the first and middle part of the day followed by thick haze during the latter part of the day. Establishing the time of day that the "First Part", "Middle Part" and "Latter Part" represented was not a trivial matter as the times were not explicitly documented within the pages of the Collection. However, some of the pre-printed Abstract Log forms had a footnote that read "Whether the day commences at noon or midnight, always call from noon to 8 P.M. 'First Part' ." This still did not clarify if the record started or ended at noon of the date entered. Since an observer's guide or handbook was not available for the period, the closest available substitute was the minutes from the 1853 International Conference in Brussels (Maury 1854). In the minutes' "Explanatory Notes for Keeping Abstract Logs" it was stated that "The direction and force of the wind should be regularly entered at 4 A.M., noon, and 8 P.M. The force and direction entered should be that which has been most prevalent during the eight preceding hours." This still did not clarify the date question. The earliest U.S. published instructions to the observer we were able to locate were printed in 1883. They contained a note saying that on September 1, 1876, work was begun on a new edition of meteorological charts of the ocean. The instructions called for observations to be taken every two hours with a summary section titled "Particulars of the Weather During the Day" which partitioned the day into three segments: midnight until 8 a.m., 8 a.m. to 4 p.m., and 4 p.m. until midnight, somewhat different than the instructions outlined in the minutes from the Brussels Conference. An example from a ship reporting on July 21, 1887, which was filled out according to the instructions, was located in the archives along with the instructions. This raised the possibility that the daily record was for the 24 hours following the date given. However, the 1883 instructions made reference to distance run since the preceding noon and the current during the past 24 hours. Based on an article by Oliver and Kington (1970) that describes that entries in the British ship logs were for the previous 24 hours and that the "First Part" covered two four hour watches (noon - 2000) and the "Middle Part" (2000-0400) two additional watches, and the "Latter Part" (0400 - noon) another two four-hour watches. British logs are a significant part of the Maury Collection and there appeared not to be any differences in reporting procedures based on country of registry. These facts in conjunction with further evaluation of the Maury logs almost assuredly indicated that the observations were for the previous 24 hours. For example, a few of the newer, circa 1870s observational forms found in the Collection were designed for reporting two hourly observations. A few of these contained entries similar to dividing the day into three parts thus providing further evidence that the general practice at the time was to provide observational information covering the previous 24 hours. Ditto marks (") were sometimes entered as two small parallel marks, but in many cases a "do" or "same" was used as a substitute or abbreviation for ditto. Recognizing ditto or the equivalent symbol or abbreviation was not much of a problem, but at times it would appear for several days and often refer to the entry to its left rather than the entry above as the daily observations were subdivided into three eight-hour periods (noon to 2000, 2000 to 0400, and 0400 to noon) representing the previous 24 hours before the ship's position was entered into the ship's log at local noon when the sun was directly overhead. It was tedious at times ensuring the right value was entered in place of the ditto mark. For example, we believe the ditto in the wind direction referred to the value to the left. A ditto in the middle section would refer to the same direction as reported for the first part of the day and a ditto in the latter part would refer to the direction reported in the middle part of the day. If a ditto appeared in the first part it would be the same as reported in the latter part from the previous day and so on if more were reported. Two elements in the Maury Collection that were digitized when available were the ocean currents and the magnetic variation, both of which often had log entries that appeared unrealistic. Neither of these two elements are part of the data reported through present ship weather reports. Contemporary ocean current observations are generally obtained from oceanographic research ships or drifting buoys which are based on consecutive reports containing time and position. Many different variations in units (e.g. knots, miles per hour, miles traveled in 24 or 21 hours, meters per second, kilometers per hour, etc.) were used to report the ocean current movement in the Maury Collection, and when the units could not be established at the time of digitization, an unknown units code was entered. This allows the information to be studied at a later time to determine if the value can be salvaged. Similarly, magnetic variation data were digitized just as the observer had entered them even if apparently erroneous, so they also can be studied in more depth at a later time. Further investigation might reveal an appropriate interpretation of this information. In one log, the entries showed the direction of the ocean currents as SW45, NW57, etc., which is not standard terminology. Our best estimate was that it should have been entered as S45W (South 45 degrees West) and N57W (North 57 degrees West). Observers were generally consistent in reporting a given element throughout a log. Almost any marine data set that has not been processed through a quality control routine will have a number of problems with ship positions. The Maury Collection had its share of records where questionable ship's positions had to be adjusted or modified during the keying process. This was generally done to make the ship's route fit the expected track based on the ship's average progress and a normal projected track between its departure port and its destination. If there were many position errors in the final database one would expect a number of positions to appear over land. This did not occur, as was shown by the locations of the digital records that had latitude/longitude position reported. Once plotted on a map, very few appear to be landlocked which is a good indication that the quality of the observation locations is relatively good. Approximately sixty-five percent of the records keyed had positions available. Often the crew was unable to obtain a position (e.g., clouds prevented a noon sun reading) and for a few voyages only the latitude was entered, never any longitudes. Some of the logs also contained dead reckoning positions. In some cases there was considerable disagreement between the observed and dead reckoning positions. In other instances the agreement was good and additional positions could be provided by using the two in combination. Even though thirty-five percent of the observations lack a position, an interpolated position can often be derived. This is scheduled to be done as the data are placed into a common database. These interpolated positions should be rather reliable as the sailing ships were relatively slow. Positions were occasionally derived during the preparatory phase of the digitizing process at CNODC from a landmark or port name entered by the observer to indicate the ships location. Occasionally, digits were inverted by the observer, or the observer was off by ten degrees or some multiple of the longitude position. These were corrected during the data preparation stage before keying, whenever the problem was identified. In general, the wind force in the Maury Collection was in descriptive terms entered in the "Remarks Section" along with a summary of the weather encountered (e.g. fine weather, squally, cloudy with rain, fair, fresh breeze with rain showers, terrific rain, brisk breeze and overcast, etc.). Most of the log entries were dated before the Beaufort force and Beaufort weather notation gained wide acceptance and usage among the world's Merchant Marines. We devised three ways of coding the present weather in the keying format: (1) the Beaufort Weather code; (2) a four-choice code which appeared on some forms allowing the observer to report fog, rain, snow, hail or any combination of the four; (3) descriptive terms that could be adapted to the modern two digit (00-99) WMO Code 4677 or Beaufort notation. The code used was documented by a leading indicator. Reference the data format for more details. In hindsight, one valuable lesson learned from this digitizing project is that there should be a special weather code for those cases when the observer reported non-significant weather such as fine weather, pleasant conditions, or good weather, when the Beaufort weather code or WMO code 4677 did not provide a match. What resulted was that nothing was digitized for these entries, leaving a user of the data to guess at the weather when a code was not keyed. It cannot be determined from the digitized data whether the present weather was not significant or simply not reported or written so poorly that it could not be read and thus digitized. Substituting a code to indicate no significant weather whenever the present weather is missing will most assuredly bias the statistics towards fair weather. In the earliest logs of the Maury Collection often the only element reported was the wind direction without any reference to the speed (force). After this earliest period most of the observations did contain a descriptive indication of the wind force in the "Remarks Section" along with wind directions reported for the three 8-hour segments of the day. These descriptive terms range from expressions that match the common Beaufort force terminology to vague descriptions that provide little information as to the wind force at the time of the observation (e.g. declining, increasing, steady). In preparing the records for digitizing, CNODC was very careful in determining the wind direction abbreviations that, if not carefully inspected, could lead to misinterpretation. Unclear penmanship could cause an "N" to look like a "W" or vice versa; S by W (south by west) would often appear as "SSW"; penmanship and the foreign language abbreviations (e.g. "O" meaning west (Ouest) in French and east in German (Osten) , Russian (Ost), or Danish added to the confusion. The Russian elite often used French during this period to prepare official reports, which often included ship logs. Reviewing the previous and following observations would often help to determine the intent of the observer in a given report. This would often involve intense discussion among several data preparers at CNODC before a consensus could be reached as to what the observer meant. Abbreviations in foreign languages also had to be reviewed carefully to ensure the proper hemisphere was assigned. Unclear handwriting was also the case where it appeared that a "V" or "W" had been entered on the log in both the wind and remarks sections, raising a question about the meaning. After close scrutiny it was determined that they were both very sloppy entries of the ditto abbreviation "do". Once this was determined, the entries made sense and the digitizing was able to proceed. A few of the logs had the wind direction encoded using "Raper's Symbol" which required some research and conjecture to determine the meaning. The four quadrants of the wind direction are represented by two perpendicular lines (a large plus sign), and a number in the appropriate quadrant indicates which quadrant the wind is coming from. The number also divides the quadrant into eight sectors (eight equal portions) in a clockwise direction, where four quadrants and eight sectors divides the wind direction into a thirty-two point scale. For example, a "6" in the upper right quadrant translates to an ENE wind direction, a "2" in the lower left quadrant indicates to a SSW wind, and a "1" in the lower right quadrant records an ExS wind. A "0" overlaying the center of the plus sign indicates calm winds, and a "V" overlaying the center indicates a variable wind. A "V" in the upper lefthand quadrant indicates variable winds from the northwest. If the winds were variable from NW by N to S by W, a "3" would be placed in the upper left quadrant and a "1" in the lower left quadrant. All these entries had to be translated to the standard thirty-two point scale. A major concern is whether the reported wind directions (those entered in the log) are true or magnetic. The only guidance available for the mid-1800s is found in the minutes from the Brussels meeting in 1853. There was some controversy at the meeting over whether the wind directions should be corrected for local variations (effects of the ship); based on the vote it was recommended that these be made but that the direction of the wind be the magnetic direction. The minutes read: " Column 7.- Direction of the Wind. The direction of the wind is the magnetic direction, with due allowance for appearance caused by the motion of the vessel. It is the direction of the wind which has prevailed for the last 8 hours. It should be expressed to the nearest point of the compass." We assume all the wind directions are magnetic directions and must be corrected for the magnetic declination to adjust them to true winds as currently reported. In the Hydrographic Office Instructions (1883) it was stated that "when the wind has any appreciable force, and can possibly be averaged for the two hours, then its mean magnetic direction to the nearest whole point for the two hours is to be recorded." It appears that it was not until 1906 (U.S. Department of Agriculture, 1906) that instructions to the observer in the first edition of the Circular M first requested true wind directions be reported. The 1906 instructions stated "The direction of the wind to be recorded is the true direction, not the magnetic. Its direction as given by the compass should therefore be corrected for the magnetic variation and for the deviation (if this is large, as is frequently the case on board iron vessels). In several instances notes were attached to Maury Collection logs stating the winds were magnetic, but none were found to indicate the wind directions were true. One unusual log (Reel Seventy-four, frame 834) contained wind force information using a code of 0-8. Based on rather general description information accompanying the code, we converted it to the standard Beaufort force as follows: Code (0-8) Conversion to Beaufort force (0-12) ----------------------- ------------------------------------ 0 calm 0 1 light airs 1 2 light breeze 2 3 moderate breeze 4 4 fresh breeze 5 5 moderate gale 7 6 fresh gale 8 7 tremendous gale 10 8 hurricane 12 Chenoweth (1996, personal communications) pointed out that since the establishment of the Beaufort Scale, even though little change has taken place in the descriptive terms, the frequency of reporting certain force numbers appears to have changed over time. He discovered this during his search through numerous ships' logs in preparation for his study of 1816, the year without a summer (Chenoweth, 1966). If one assumes that the wind climatology remains relatively stable across various geographical regions one would expect the force frequencies to remain rather constant over time. However, he noticed the frequencies differing with time and assumes that the mariner's reporting practices or their perspective of a given force changed over time thus introducing a change to the mean wind speed if based on the force numbers as reported. This possibility will have to be investigated. The barometric pressures in the Maury Collection were keyed as entered in the logs without making any adjustments; in general the logs carried no indication of whether barometric entries were "corrected" or "as read". It is unclear, therefore, if they were corrected for instrument error and, in the reports where a mercurial barometer was used, whether they were corrected for temperature and gravity. It was interesting to note that in the instructions to the observer printed in 1883 under the section covering barometers, the following comments on attached thermometers was included: "A thermometer is sometimes attached to the aneroid, as a convenience for obtaining the temperature of the air. As regards the instrument itself, no correction for temperature can be applied with certainty." It was not typical to find attached thermometer entries in the Maury Collection although there was generally a column titled "Ther. Att'd." However, in many of the early twentieth century reports, after the Maury Collection, the information was included along with an entry indicating the type of barometer, mercurial or aneroid. In a few cases, however, sufficient metadata are available to allow confidence that the pressure entries were probably corrected to sea level. For example, the 1856 log from the ship "S.H.Talbot", which sailed from Richmond, Virginia, to Melbourne, Australia, with a load of flour, documented the height of the ship's mercurial barometer (eighteen feet above sea level) and included entries at the top of the page providing clues as to the general practice aboard at least this particular vessel. The two notes were "corrected before writing in columns" and "this journal is kept in sea time". This would lead us to believe that the pressure entries were probably corrected ones, at least for this log. Another odd circumstance was that the first observation entered in the log was after they had been to sea for ten days. From that first delayed entry forward, the weather observations were consistently entered until reaching Melbourne where a note was included stating "99 days from Richmond to Melbourne". For one voyage CNODC was preparing to digitize, it was noted that the pressure values in inches were much too low to be in English units. In trying to resolve the problem it was noted that above each pressure column the word "Paris" had been entered. A literature review indicated that during this period mercurial barometers built in France were based on the "Paris Inch." An article (Lamb, 1986) describing past units documented the conversion factor from Paris to English inches. After applying this correction the reported values fell within expected limits for the given locations. Each Paris inch was subdivided into twelve parts called lines, and each line in turn was subdivided into tenths. That is, pressure was reported at the resolution of tenths of a line, with e.g., 11.9 lines being the largest value reported before advancing to the next whole unit of inches. For example 27'7.6 would represent 27 inches plus 7.6 lines of an additional Paris inch. To convert to English inches required for the keying format, the 7.6 would be divided by 12 to get units to hundredths of a Paris inch (7.6/12 = .63) then 27.63 Paris inches would be multiplied by 1.066 to convert to English units: 27.63x1.066 = 29.45 English inches). The marine observers would also forget at times to change the hemisphere indicator when the ship crossed the equator, Greenwich Meridian, or international date line, so the CNODC data preparers had to be very careful to pick up most of these errors. There were cases where critical date information seemed to be missing. In one instance the voyage had been copied in reverse sequence on the microfilm and in others the year of the voyage would only appear once in the logbook, making it difficult to locate. Such information could have been missing at CNODC if the microfilm image was so weak that the frame had not been copied and forwarded to China. In some instances the observer would use Roman numerals to indicate the month. Variations of this nature appeared throughout the logbooks in the Collection, thus significantly complicating digitization. Interesting Remarks and Added Information: In addition to simply reporting the weather conditions observed, other entries were often placed in the logbooks, providing some insight into the life of a mariner of the day and of potential interest to historians as well as meteorologists. Many of the remarks referred to the number of sails that were being carried, which could provide some correlation to the wind force reported. Unfortunately, it was not practical to digitize these or the other general remarks because of the non-standardization and difficulty in reading the penmanship. Occasionally entries would appear if the crew were suffering from illness or if a sailor had died aboard. There were rare comments about the poor food or the poor performance of the crew. Occasionally there were entries on sightings of whales, birds, landmarks assisting navigation, or the passage of another ship which they signaled. One well-documented voyage recorded included remarks about an encounter with a tropical storm. These remarks included praying to survive the storm and sightings of human bodies from a nearby island floating in the sea along with their livestock that had not survived the fury of the storm. Another interesting log, from the ship "Montreal" sailing from New York to Vera Cruz in 1848, contained the description of a drunken crew and the difficulty they had getting 6 of the twelve on deck to stow the anchor and clear the decks for departure. Days later an entry describes one crew member, still sick, "out of his mind through intemperance." So it would appear that the logbooks in the Maury Collection might be of general interest to others in addition to meteorologists. There were a number of logs from whaling ships where they sailed between the coasts of South America and Australia and as far north as Hawaii for months on end. During the 1830's to 1850's it was not uncommon for a ship to spot as many as seven pods of whales within a month and take as many as five individual whales, sometimes two from a single sighting. Caution: If a "02" appears in the Form Type field (positions 48-49) in the header, the observations are hourly versus daily and often the incorrect date (off by one day) was digitized. Also during the digitization, the hours were often not converted to a twenty-four clock and there is no indication of whether the time is A.M. or P.M.. The problem of incorrect day/time will be corrected using record sequence and comparisons against the microfilm archive during the conversion to the COADS common format. References: Bates, C.C., and J.F. Fuller, 1986: "America's Weather Warriors, 1814-1985", Texas A & M University Press, College Station, 360 pp. Chenoweth, M., 1996: Ships' Logbooks and the The Year without a Summer'. "Bulletin of the American Meteorological Society", Vol. 77, No. 9, 2077-2093. Fletcher, J., U. Radok, and R. Slutz, 1981: Historical ocean climate project. "Trop. Ocean-Atmos. Newslett.", No. 5, 1-11 Hawthorne, H., 1943:"Matthew Fontaine Maury, Trail maker of the seas". Longmans, Green and Co., New York and Toronto. 226 pp. Hydrographic Office, 1883: "Meteorological Journal Instructions", Government Printing Office, Washington. Kalnay, E. Et al., 1996: The NCEP/NCAR 40-year reanalysis project. "Bull.Amer. Meteor. Soc.", 77, 437-471. Lamb, H.H., 1986: Ancient Units Used by the Pioneers of Meteorological Measurements, "Weather", Vol. 41, No. 7, 230-233. Lewis, J.M., 1996: Winds Over the World Sea: Maury and Koppen. "Bull. Amer. Meteor. Soc.",77, 935-952. Lewis, C.L., 1927:" Matthew Fontaine Maury, the pathfinder of the seas". The United States Naval Institute, Annapolis, MD, 264 pp. Maury, M.F., 1854: "Explanations and Sailing Directions to Accompany the Wind and Current Charts", Sixth Edition, Philadelphia, 772 pp. NARA (National Archives and Records Administration), 1981: The Maury Abstract Logs, 1796-1861. (Microfilm Publication M1160; Records of the Weather Bureau; RG 27), NARA, Washington, D.C. NCDC (National Climatic Data Center), 1968: TDF-11 reference manual, NCDC, Asheville, NC. Slutz, R.J., S.J. Lubker, J.D. Hiscox, S.D. Woodruff, R.L. Jenne, D.H. Joseph, P.M. Steurer, and J.D. Elms, 1985: Comprehensive Ocean-Atmosphere Data Set; Release 1. NOAA Environmental Research Laboratories, Boulder, Colo., 268 pp. (NTIS PB86-105723). U.S. Department of Agriculture, Weather Bureau, 1906: "Instructions to the Marine Meteorological Observer of the U.S. Weather Bureau, Circular M - 1st edition", Government Printing Office, Washington. U.S. Department of Agriculture, 1938: " Atlas of Climatic Charts of the Oceans", Weather Bureau, W.B. No. 1247, Washington, D.C. Williams, F. L., 1963: " Matthew Fontaine Maury, scientist of the sea". Rutgers University Press, New Brunswick. 720 pp. Woodruff, S.D., R.J. Slutz, R.L. Jenne, and P.M. Steurer, 1987: "A comprehensive ocean-atmosphere data set". Bull. Amer. Meteor. Soc., 68, 1239-1250. Woodruff, S.D., S.J. Lubker, K. Wolter, S.J. Worley, and J.D. Elms, 1993: "Comprehensive Ocean-Atmosphere Data Set (COADS) Release 1a: 1980-92". "Earth System Monitor", 4, No.1, 1-8. Woodruff, S.D., H.F. Diaz, J.D. Elms, and S.J. Worley, 1997: COADS Release 2 Data and Metadata Enhancements for Improvements of Marine Surface Flux Fields, European Geophysical Society (EGS) XXII General Assembly, Vienna, Austria, 21-25 April, 1997, "Proceedings Journal Physics and Chemistry of the Earth (and Solar System)" (submitted).