John Philip was struck by a car and killed on Saturday 26 June 1999 in Amsterdam where he was visiting the Centre for Mathematics and Information Science. He was a Fellow of the Royal Society, a Fellow of the Australian Academy of Science, a Fellow of the American Geophysical Union, a Foreign Member of the All-Union (later Russian) Academy of Agricultural Sciences, and only the second Australian Foreign Associate of the US National Academy of Engineering. He was the first non-American recipient of the Robert E. Horton Medal, the highest award for hydrology of the American Geophysical Union. In 1998 he was made an Officer of the Order of Australia for ‘service to the science of hydrology, to scientific communication in promoting the interests of science for the community, and to Australian culture through architecture and literature’. This memoir discusses John Philip’s character and his work as Australia’s most distinguished environmental physicist. It explores his management of science and his role in the Australian Academy of Science as well as his poetry and his fascination with architecture.
John Philip’s father, Percy, was a farmer from Franklinford, near Castlemaine in Victoria, who moved to Foster after his marriage and became a stock and dairy inspector. John’s mother Ruth (née Osborne) was a schoolteacher and a Methodist lay preacher who had a deep commitment to education and very high expectations for John and his younger brother, Graeme. With his mother’s encouragement, John duly developed a love of learning, a fund of biblical quotations, a smattering of Greek, and snatches of hymns with which he would irritate audiences in later years. Under his mother’s influence, he developed a precocious mathematical talent and he won an open scholarship to Scotch College in Melbourne at the early age of eleven. The family then moved to Carnegie on the outskirts of Melbourne so that John could attend Scotch as a day boy. At the age of twelve, John demonstrated his independence by walking out of an evangelical service when asked to promise his life to Christ. Frances (Fay) Julia Long from the girls’ side of the assembly was his sole fellow dissident. They were married ten years later.
At Scotch College, John was placed in a class where he was almost three years younger than his fellow students and where his intellectual world expanded enormously. He also discovered that his English teacher would accept poetry in place of an essay; his lifelong love of poetry was born, he said, from his discovery that poetry brought the greatest results for the least effort. He matriculated at 13 and spent a further two years studying ‘leaving Honours’ before he could enter the University of Melbourne. John said later that mathematics was his favourite subject at Scotch although he was not identified as exceptionally talented. This was probably because, in his class of six students, R. H. Dalitz, A. K. Head and John himself became Fellows of the Royal Society, J. B. Swan and S. N. Milford became professors of physics and N. D. Symonds became a biophysicist who later worked with Max Delbrück and Erwin Schrödinger.
John entered Queen’s College in the University of Melbourne to study Engineering in 1943. There, a colleague recalled, his life was characterized by brief periods of intense activity interspersed with indolence and Frances seemed the one fixed point in his restless world. Neither the Master of Queen’s nor G. H. Vasey, his academic supervisor, identified him as an outstanding student. Vasey, however, warmly supported his early appointments because of his originality, his ability to apply himself and to work extremely hard in areas that attracted his interest, and his above-average competence in English usage. Furthermore, at 19 years of age, he was, and remains, the youngest Civil Engineer ever to graduate from the University of Melbourne. John later asserted that, at graduation, the message he took away from the course was that ‘all things are understood, and all a young engineer needs to know is what handbook to use’.
When he graduated, John was too young to be paid the Victorian Public Service Engineer’s adult wage but the University, for the first time ever, advertised for a graduate assistant in agricultural engineering at adult rates. He was appointed and seconded to the Irrigation Research Station of the forerunner of CSIRO, the Council for Scientific and Industrial Research at Griffith, New South Wales. It was a revelation to John that, for agricultural scientists struggling with the hydraulics of furrow irrigation, all things were not understood and no handbook existed. With his acute mathematical and physical insights he quickly identified a range of problems concerning water movement in the soil–plant–atmosphere environment that provided the scientific focus and sense of purpose that had previously eluded him. His original approach to these problems, his engineering aptitude, and his newly discovered enthusiasm to apply it prompted a comment from Vasey that ‘when he left Griffith he left in the minds of the Extension Organization and quite a number of farmers a regret that the service did not employ a full time engineer of the Philip type’. John, for his part, recalled that ‘I blundered into a vocation that turned out over the past 50 years to be more fun than work’.
In 1948, John joined the Queensland Water Supply Commission with responsibilities for design in the Burdekin and Mareeba Irrigation Schemes. His supervisor, T. A. Lang, in his later role of Associate Commissioner with the Snowy Mountains Authority, recorded that during this time John was ‘faced with an almost entire lack of basic information and exhibited considerable intelligence and aptitude in handling irrigation problems’. He also commented: ‘He is apt to be a little untidy in his appearance. This may be the result of his hobby, which is writing poetry and his tendency toward a certain Bohemian outlook in his private life. This in no way affects his work which is technically of a high standard’. Lang was referring to John’s link, through the magazine Barjai, to Brisbane bohemia, including the artist Charles Blackman, the poet Barrie Reid, and Charles Osborne, now a London music critic and writer. Vida Horn, another member of the group, remembered John as ‘possessing a maturity at age 20 most of us lacked. It was as if his life was precisely laid out and he was in charge’. She said that she knew of ‘no other poet who could transmute the language of science to that of poetry with such simplicity and elegance’.
In 1949, while working in Brisbane, John married Frances in Queen’s College Chapel at the University of Melbourne. Frances’s practical and sweet nature together with her artistic ability contrasted with and complemented John’s personality. As John said in his retirement speech in 1992: ‘My greatest piece of luck is to have fallen in love with someone of such keen intelligence, lively wit, independence of spirit, and downright courage. I cannot imagine anyone who could have helped me so much, or kept me on the rails so well’.
After his experiences in Griffith where he had had free intellectual rein, John became increasingly unhappy working within the constraints of a state engineering bureaucracy, however. The general ignorance of the theory behind his work worried him and he yearned for the freer and more creative environment of CSIRO where he could deal with more basic issues of science. In his words ‘liberation came’ with a research appointment in 1951 to the Regional Pastoral Laboratory of the CSIRO Division of Plant Industry at Deniliquin, New South Wales. With post-war housing still in very short supply, John and Frances spent the summer at the end of 1951 living in a tent in an orchard at the edge of Deniliquin. This accommodation was described in official correspondence as ‘so primitive that the arrangements had to be abandoned’ and they were moved to the Royal Hotel. That was also temporary and, by June 1952, with Frances seven months pregnant, it was suggested by the Officer-in-Charge at Deniliquin that the post-mortem room at the town’s ancillary hospital be converted to a temporary residence for the young family. John later asserted that, throughout this period, he was so enchanted with the freedom to be creative that he was quite unaware of these privations although he disliked camping and barbecues to the end of his life.
In Deniliquin, the family developed a pattern of behaviour where Frances, at the cost of her career as a talented painter, took full responsibility for their three young children and maintained a stable home so that John was completely free to pursue his career. This freedom of action, coupled with his extraordinary ability to discern fundamental physical elements of an environmental problem, to formulate the problem mathematically, and to focus huge energy to find practicable solutions were the keys to his genius.
Shortly after John’s appointment, Sir Otto Frankel became Chief of the Division of Plant Industry. Frankel identified with the research ethos of Sir David Rivett’s CSIRO, ‘to find the best person for the task and give them the freedom to get on with it’, but he was ill at ease with John’s mathematical and physical approach to environmental problems. Professors Pat Moran and John Jaeger at the Australian National University reassured him, however, and John Philip followed his scientific instincts. John regarded Jaeger as the closest person he had to a scientific mentor and was delighted to receive the Jaeger Medal of the Australian Academy of Science in April 1999.
John’s intellectual associations in Deniliquin were strengthened in 1956 by the appointment of Dan de Vries to Deniliquin. Dan had just completed his doctorate at Leiden based on research in the physics laboratory of what is now Wageningen University and his strong physical sciences background, his enthusiasm to apply physics to real-world problems and his sound experimental skills complemented and extended John’s horizons.
John Philip’s Deniliquin years were enormously productive and from 1953 until 1960 he published more than forty scientific papers, although his work habits, as in his undergraduate days, still varied between periods of intense activity when he often worked all night and periods when he was a quite disruptive influence in the laboratory. Nevertheless, he was gaining an international reputation. His interests ranged from population dynamics to heat and mass transfer in the biosphere, and a scan of his first ten years of publications reveals how catholic his commitments were. A critical step came with a visit in 1956 to Dr E. C. Childs at the Agricultural Research Council Unit of Soil Physics in Cambridge. Ernest Childs’ ‘clear thinking and decent rigor in a field where scientific standards have seemed, all too often, to have received little consideration’ strongly influenced him, as he later acknowledged.
On his return to Australia, John’s reputation, as well as that of de Vries, was established by the joint paper (14) on heat and mass transfer in unsaturated soil that won the Horton Medal of the American Geophysical Union in 1958. A brief visit to the California Institute of Technology also resulted in an effusive letter from Professor James Bonner to Otto Frankel with an invitation for John to spend sabbatical leave at Caltech. This proposal confirmed in CSIRO that, while he had a difficult personality, John was very able and an asset to be nurtured. Otto therefore agreed that, on their return from Caltech, the Philip family should move to Canberra where John would establish an Agricultural Physics Section in the Division of Plant Industry.
Following Otto’s advice to forgo a PhD, John took a DSc (physics) from the University of Melbourne in 1959 for the extraordinary set of papers (5, 12, 16, 17, 19–21, 29) that brought unity to the existing approaches to water movement in soil.
John’s reputation, as a very able young man in a hurry, was growing. In 1963, when John Falk succeeded Otto Frankel as Chief of the Division of Plant Industry, John became one of his four Assistant Chiefs and, in 1967, at the comparatively early age of 40, he was elected to the Australian Academy of Science. Then, in 1969 when John Falk became very ill, John became Acting Chief of Division. In 1971, however, Lloyd Evans was appointed substantive Chief and John became Chief of a new, small and autonomous Division of Environmental Mechanics. The new Division’s objectives emerged from those of the Agricultural Physics Section and sought to link laboratory experiments with field behaviour and to develop practical mathematical descriptions of environmental processes. The Division was created about three small groups set up to investigate and bring together the components of the soil–plant–atmosphere continuum that was conceived by Gradmann and van der Honert to unify the terrestrial hydrological cycle. This concept (15) recognised that water in the soil, the plant and the atmosphere forms a thermodynamic continuum. Water flows from one domain to the next along gradients of water potential, so its flow could be analysed in a mathematical–physical framework. Coupled with the concept of a critical water potential at which plants lose turgor and stomata begin to close, the analysis could be used to predict how the properties of each domain controlled transpiration and water extraction by plants and the onset of wilting. A fourth group, called Applied Mechanics, which John himself led, provided theory complementing the three more experimental groups. John insisted on scientific quality and his support for a series of very distinguished Pye Fellows ensured that Environmental Mechanics was recognised internationally as a centre of excellence. Except for a three-year period as Director of the CSIRO Institute of Physical Sciences, John was Chief until his retirement in 1992. He then became the first CSIRO Fellow Emeritus.
John’s retirement saw no diminution in his research. He continued to collaborate internationally and delivered his last paper in Amsterdam, two days before his death.
Modern theories of mass and energy movement in the biosphere, focused on water, were generally accepted by the mid-twentieth century. They tended to be reductionist in character and flow equations combined macroscopic material, force and energy-balance equations with flux laws based on space gradients of potential. These equations were difficult to solve because the transfer coefficients tended to be strongly related to the local concentration of the entity of concern, the location, or both. The architecture of the crop canopy and the root system complicated their formulation as well as the scale of their application and test. Nevertheless, their solutions were required to deal with important problems of land and water management and crop and forest production. When computers were in their infancy, John sought practical methods for description and measurement in each phase of the soil–plant–atmosphere continuum but his principal interest was in soil water and his initial focus in Deniliquin in 1951 was on border and furrow irrigation.
John concentrated on the Lewis–Milne equation, which had been developed in 1938 to describe the advance of an irrigation front across the soil surface and which requires an explicit description of local infiltration of water into the ponded soil behind the advancing front. Infiltration equations then in use were largely empirical and John’s first journal paper on the topic (2) analysed border irrigation using such equations. At the same time, he sought more physically based formulations and, in the process, rederived an infiltration equation of Green and Ampt (1911), Australia’s first soil physicists, although he was unable to use it for the border irrigation problem because it does not calculate the infiltration rate explicitly. He therefore returned to the basics of infiltration theory and focused on the Richards equation.
L. A. Richards (1931) had formulated a general flow equation for water in unsaturated soil, although, as John later said, ‘for more than 20 years, Richards’ equation lay around like some strange object fallen from the sky. The natives looked at it with some awe, but knew not what to do with it’. It combined material balance for the water with Darcy’s law applied to unsteady water flow in unsaturated soil and, by 1951, it was generally recognised that this equation could be written, in John Philip’s terminology, as
In these equations q is the volumetric soil water content, t is time, Ñ is the vector differential operator, K is the water-content-dependent hydraulic conductivity, z is the vertical co-ordinate, and y is, essentially, Edgar Buckingham’s (1907) capillary potential of the water.
The first term on the right-hand side in these equations represents water flow due to capillary and surface forces; the second term describes the effects of gravity. The solution of these equations is complicated because of the gravity term and also because both K and q are material characteristics that depend strongly on y. Furthermore, q (y) is hysteretic.
John Philip brought order to the solution of these equations. He used the Childs and Collis-George (1948) definition of a soil water diffusivity, D(q) = K dq/dy, in equation (1) and improved an approach of Arnold Klute in his PhD studies at Cornell. Klute’s (1952) approach ignored the gravity term on the right-hand side so, for non-hysteretic flow, equation (1) becomes a non-linear diffusion equation describing flow due only to capillarity. John developed an iterative, quasi-analytical solution for this equation for the particular case of a step change in water content at one end of a long horizontal uniform column of dry soil. His approach converged rapidly and was more general than that of Klute. It was also agnostic about the forms that D(q) might take and required only that the water retention and hydraulic conductivity characteristics of the soil existed and were measurable. He then included gravity by formulating a series solution for equation (1) in powers of t1/2. This followed his realisation that his gravity-free solution could be considered the first term in such a series for one-dimensional flow and that the effect of gravity could be seen as a perturbation on this solution represented by subsequent terms in the series. The coefficients of the terms were the solutions of linear equations and the series converged rapidly. John also perceived that a travelling wave represented the long-term asymptote of his series. He accomplished this work during his early days at Deniliquin and reported it in the papers for which he was awarded his science doctorate.
This analysis was mathematically and physically novel and it identified general patterns of behaviour that are now recognised and used across all manner of systems. An early outcome, for example, was the recognition that during the ‘initial’ stages of infiltration, when the diffusion equation appears to prevail, cumulative infiltration is proportional to t1/2 with the slope of this line a characteristic of the material and experimental conditions. John termed this characteristic the sorptivity (20). It predicts the initial response of soils to rain or irrigation and, because it is readily measured, it is often used to infer the K(y) characteristic of porous media.
These insights for one-dimensional flow with concentration boundary conditions were extended to multidimensional flow in his ‘Theory of infiltration’ of 1969 (89). This citation classic explored not only the origins and the formal solution of the basic equations but also discussed operationally useful approximate solutions and the limits to their application. John’s illustrative calculations were based on experimental data of Moore (1939) for a soil identified as a Yolo light clay and his intuitive extension of Moore’s data to very low water contents led him to anticipate that a total diffusivity of the soil water would have a minimum in a region of transition from predominant movement in the vapour phase to predominant movement in the liquid phase. This phenomenon has since been verified experimentally for a wide variety of soils and other porous media.
He also illustrated how material properties permit estimation of the time before the diffusion approximation becomes inappropriate in one-dimensional (vertical) flow, and when a one-dimensional approximation ceases to be appropriate for two- and three-dimensional flow. These general insights are conceptually and practically beneficial although they are obscured by brute-force computer solution of the Richards equation that characterizes modern modelling.
The behaviour of swelling soils and other porous media was also amenable to this theory. John Philip’s interest in these materials arose in 1967 when he was asked his opinion of filtration experiments on swelling clay that had been analysed using solid-based space-like co-ordinates. Within a week, John had asserted that there was no future in the use of material co-ordinates for this class of problem and had re-analysed the experiments in physical space and time although he expressed his ‘considerable debt to Dr Smiles and Miss Rosenthal for interesting me in this topic’. His paper (90) on the physics and mechanics of the problem remains a paradigm of clear analysis and is honoured by at least one verbatim, but unacknowledged, reiteration in the chemical engineering literature. Papers by Pieter Raats and Klute (1968) from the University of Illinois, however, forced him to revise his opinion of the use of material co-ordinates. He also realized that use of material co-ordinates resulted in an equation of exactly the same form as that of Richards for water flow in non-swelling materials so he could immediately apply his one-dimensional theory of infiltration of water in unsaturated non-swelling soils to many swelling systems. A flood of papers on equilibrium and flow in swelling materials followed (eg, 111).
In swelling systems, too, vertical displacement of wet soil accompanies water content change and, where the solid specific gravity is greater than one, there results an increase in the gravitational potential energy of the system during infiltration. This contrasts with the decrease that is observed in non-swelling soil, so vertical infiltration in these materials is analogous to capillary rise in non-swelling ones. The magnitude of this effect is moderated by the volume–water content–load relationship of the swelling system. John’s analysis of the phenomenon relied on the work of Croney and Coleman of the British Road Research Laboratories and is operationally and mathematically practical although not formally exact (107). Nevertheless, he delighted in this significant bouleversement, as he called it, of conventional wisdom and it provided many free lunches.
At this time, too, he and John Knight (120) improved Yves Parlange’s (1971) novel solution of the Richards equation. John’s disparaging treatment of Yves’ paper, however, resulted in more than a decade of ill feeling. The simple, rapidly converging Philip/Knight amendment extended analyses of equilibrium and flow during filtration, sedimentation and centrifugation in chemical engineering. He also explored the nature of stress in colloidal suspensions and applied the approach to systems where particle-to-particle interaction occurs ‘at a distance’ (101).
Hydrodynamic dispersion and chemical reaction studies in soils followed a visit to David Elrick of the University of Guelph. The common approach then, based on the Saffman and Taylor theory of flow of solution in capillary tubes, was to perform ‘breakthrough’ experiments where emerging solute concentrations were measured following steady flow of solution through saturated soil. These integral measurements, however, gave little information about transient behaviour. John Philip and John Knight realized that transfer of solute in soil during absorption of solution by dry soil would permit analysis of hydrodynamic dispersion during unsteady, unsaturated soil-water movement (152) and this led to analysis of dispersion and chemical reaction during unsaturated water flow in non-swelling and swelling materials.
John’s recognition of an analogy between the mathematical physics of light scattering and quasi-linear water infiltration in soil provides another example of the way he considered problems. His interest was attracted by the work of Trevor Waechter and, with Waechter and Knight (222), he illustrated the analogy in studies of the ‘watertightness’ of cavities designed to store noxious materials in essentially unsaturated soils.
While John Philip’s principal interests lay in water movement in porous materials, his experiences with Dan de Vries in the semi-arid climate of Deniliquin led them to study advection, the horizontal transport of heat and moisture due to changes in surface wetness, and the way in which soil and atmosphere interact to control evaporation from soils and plants.
Both Philip and de Vries rejected what John called ‘flat-earth micrometeorology’ where the surface is a homogeneous, semi-infinite plane and steady flux-gradient relationships in the vertical dimension are the focus of analysis. Instead, they tackled, head-on, the complexities that arise from the surface heterogeneity of agricultural landscapes, which are associated with limited fetches and sharp contrasts at boundaries and which give rise to evolving concentration fields and surface fluxes downwind of boundaries. As he said in his ‘theory of local advection’ (38), ‘In this real world, irrigated fields adjoin deserts, reservoirs are of finite extent, dry lands exist beside seas, and cornfields beside close-grazed pasture’. This work extended an analysis of de Vries (1957) by describing the vertical profiles of both the wind speed and the eddy diffusivity by power laws and solving numerically the two-dimensional atmospheric diffusion equation, subject to surface radiation, concentration and flux conditions downwind of a change in surface properties. He was concerned particularly with the partitioning of solar energy downwind of a change in surface wetness. This arose in relation to the influence of the size and position of an irrigated field within a more extensive dry area, on evapotranspiration. Frank Bradley and Norman Rider (52) tested the approach downwind of the junction between tarmac and grass in landmark experiments conducted at Canberra airport.
Problems related to surface roughness and the scales of application were evident at the time. Surface roughness effects, in particular, could be large, especially close to the leading edge, and drag plate technology developed by Frank Bradley to measure them became the standard against which other methods were assessed. John worked on other problems later. They included diffusion across the upwind edge, boundary-layer development and blending heights for checkerboard patterns, where the wind blows across many alternating surfaces with different properties (eg, 291). This work increasingly is realising its potential, and advection-type solutions are sought, for example, in long-term flux measurements where fetches are changeable, in scaling fluxes from local to regional scale where blending heights must be defined, and in dispersal of pollutants from small source areas.
John was also interested in ways to describe foliage distribution in plant stands, their light climate (70) and interactions between canopy geometry and the distributions of sources and sinks for heat, water vapour and carbon dioxide (73). In the process, he challenged John Monteith’s approach to diffusive resistances in the biosphere (129). Monteith had suggested that micrometeorological measurements could be used to infer a crop resistance. John Philip felt that the simplification of canopy exchange processes inherent in Monteith’s one-dimensional, ‘big leaf’ model was unrealistic and misleading. John Philip never published a promised detailed critique of the concept but his vehemence in attacking it led to a rejoinder in Monteith’s (1973) Principles of Environmental Physics that not only did the experimental evidence support the use of crop or surface resistance as an index of the physiological control of water loss by a crop canopy, but also no more appropriate index had yet been devised despite attacks based on armchair speculation divorced from field experience.
John Philip extended this concept to include simultaneous transfers of energy and heat (73) and his analysis of water transfer helped explain why transpiration could be restricted and plants might wilt over a wide range of soil moisture contents depending on root density, the soil hydraulic properties and the evaporative demand of the atmosphere. These ideas set the scene for a dynamic approach to plant–water relations. John’s treatment of evaporation from bare soil (18) complemented these studies and provided a physical explanation for the different phases observed in the drying of initially wet soil profiles. This phenomenon is characterized by a constant-rate phase in which the evaporation rate is that from a saturated surface and is determined only by atmospheric conditions, and a falling-rate phase that is controlled by the hydraulic properties of the soil and is essentially independent of atmospheric conditions. John’s simple either/or description of evaporation from drying soils allows relatively easy parameterization of the time course of soil evaporation from field experiments and provides for simple and robust modelling of the soil water balance.
John Philip was not good with his hands but he attached great importance to good measurement. The strong interaction in his Division of Environmental Mechanics between theory and field measurement reflected this view, and he invested substantially in methods of measurement in the very messy context of the biosphere. Personally, he analysed and proposed methodologies of soil and micrometeorological measurement and, with Dan de Vries, he explored ways to correctly measure soil heat flux, heat storage in the soil above the flux meter and evaporation-induced transfer of heat by vapour and mass flow of water (194). His last journey was to include a visit to Gerard Kluitenberg in Kansas after visiting Amsterdam, to continue work on errors in thermal conductivity probes in heterogeneous soils (304).
John Philip was alert to opportunities to develop mathematical methods throughout his studies of the natural environment and, characteristically, he used a continuum approach at a scale that he believed was appropriate to the practical problem of concern. This required that his analyses be based on macroscopically measurable average values of the entities of concern. At the same time he was aware that there might be no simple correspondence between these values and those at other scales and he explored ways to justify the form of the macroscopic and phenomenological equations from considerations of lesser scales. In his soil-water studies, for example, he sought to relate flow at the macroscopic scale to that at a pore scale. His interest in the foundation of the Richards equation led him to explore the basis of Darcy’s law in the linearity of the pore-scale Navier–Stokes equation in the limit of zero Reynolds’ number, and with sufficient homogeneity to allow averaging (13). He recognized the problem posed by the liquid–solid or liquid–gas boundary at the fluid interface (22) and later found a justification for treating such interfaces as if they were rigid (118).
He explored implications of energy dissipation associated with flow in porous media (41) and the definition of absolute, rather than the familiar differential, thermodynamic functions in soil-water studies (42). He sought to unify notions of capillary condensation and adsorption of water at surfaces (156) and was also interested in flow and transport in aggregated and heterogeneous media (196).
His mathematical methods originated largely in nineteenth-century classical physics and his style resembles that of G. I. Taylor, except that John was entirely mathematically inclined and, unlike Taylor, experimentally inept. His analysis of divergent–convergent flow in a plane region bounded by a circular wall from a point source in the wall to a point sink located diametrically opposite (91) exemplifies this approach. He showed that a solution first given by Lord Rayleigh in 1893 implied that this flow, in distinction to flow at the Darcy scale, deviates strongly from Poiseuille flow and, in particular, that on the pore scale, there is no one-to-one correspondence between the potential gradient and the flow direction.
Another example is his illustration (53) that the Taylor–Aris expression for the longitudinal dispersion coefficient associated with laminar flow in a tube corresponds to retention of only the leading term of an Eigenfunction expansion of the full solution. He also showed that, for periodic systems, the dispersion coefficient becomes a complex function of the frequency. Motivated by the problem of clogging of porous media during liquid flow, John later extended the analysis to situations where a dispersing substance is adsorbed by the wall (280). Other problems pioneered by G. I. Taylor included instability of displacement fronts in porous media (139).
In his development of analytical solutions of equations describing flow and transport, he tended to shy away from numerical solutions, although an exception was his early work on the numerical solutions of equations of the diffusion type with concentration-dependent diffusivity (5). He also, later, sometimes resorted to such solutions after reducing the problem to a simpler, more transparent form. He pursued a wide variety of methods: similarity solutions, asymptotic time-invariant travelling waves, linearization by appropriate specialization and/or transformation. This often meant solving rather complicated differential equations. In this context he was obliged to explore a wide variety of sometimes rather exotic mathematical functions, and this led to several contributions to the mathematical literature as by-products. His early work on concentration-dependent diffusion, for example, led him to a detailed study of the inverse error function and of its derivatives and integrals (46). A study of diurnal cycles in the lower atmosphere led him to resolve some problems with the definition and tabulation of Kelvin phase functions. He studied the convergence and partial convergence of alternating series and this in turn led to a more elegant and accurate form of the 250-year-old Euler–Maclaurin summation formula (204).
John’s lectures to applied mathematicians (eg, 411) tended to emphasize that he was an applied physicist and his excursions into mathematical details were always related to concrete problems in environmental physics, in line with his pragmatic approach to science and life in general. He teased mathematicians for being too pure and in the process sometimes annoyed them. As an invited speaker at an international meeting on free boundary problems, for example, he pointed out that the mathematicians’ emphasis on free boundaries arose from their fascination with generalized functions. He then showed for a number of cases that, physically, this meant emphasizing special cases, i.e. he demonstrated that mathematical generality in such cases demanded physical peculiarity! His strong feeling on this originated from his early experience reconciling the Green and Ampt type of free surface models with the Richards equation. Nevertheless, John Philip recognized very early that a free surface does arise in connection with infiltration if the diffusivity corresponding to the initial water content vanishes (17, 21). His last lecture, on 24 June 1999 in Amsterdam, dealt with a related problem in two-phase flow (306).
John enjoyed and carefully nurtured his contacts with practically minded mathematicians. Sometimes he stimulated mathematicians to work on practical problems, and he would regularly get new ideas from his contacts with them. Typically he would work out in detail or particularize their ideas and would himself, or have others do, concrete calculations. He realized that the focus on the particular would not necessarily impress the mathematicians, but would certainly be appreciated by other colleagues interested in concrete physical applications.
Applied mathematicians appear to have been most impressed by two original contributions to non-linear diffusion theory. First, he identified a large class of diffusivity functions that lead to exact solutions of the non-linear diffusion equation subject to certain boundary conditions (45). Second, he provided a detailed study of n‑diffusion, that is, diffusion where the diffusion coefficient is proportional to some power of the concentration gradient (51); this study was motivated by his interest in unsteady turbulent vertical heat transfer from a horizontal surface by free convection and in unsteady turbulent flow of a liquid with a free surface over a plane.
In summary, John Philip was a highly original applied mathematician. He developed his skills in a period when numerical models were clumsy and when his almost unique ability to relate quasi-analytical mathematical analysis to physical reality provided insights into physical processes that remain central to discriminating analysis of real-world problems. His contributions to applied mathematics were significant but his overwhelming contribution remains his ability to provide a mathematical framework to generalize limited but systematic physical observations. In this regard his ability to pick up, and run with, good experimental data derived by others was a source of great wonder and greater irritation, and his ability to expose systematic and characteristic behaviour that no amount of experiment or numerical analysis could reveal was very special.
John Philip’s first excursion into policy and management was as secretary of a committee set up to recommend ways to manage hydrology research in CSIRO. This committee, chaired by Professor E. Sherborn Hills of the University of Melbourne, recommended, in 1953, the formation of a Section of Hydrology that brought together the scientific components of the terrestrial hydrologic cycle. The proposal threatened territory of the CSIRO Chiefs and John was disillusioned when the Executive did nothing but agree ‘to record that there was some support for the development of hydrology as a science in its own right both in its pure and applied aspects and that there is both support and opposition to the establishment of a separate hydrology group either in CSIRO alone or in collaboration with a university’.
John also chaired the Science Task Force of the ‘Coombs’ Royal Commission on Australian Government Administration in 1975. The Rivett ethos strongly influenced his report, which argued for government science characterized by freedom of action, accountability to taxpayers and strong links with users. The report also argued that, because science was a central tool of policy across many departments of state, scientific strength should be established where it was to be used and facilitated by flexibility of employment and mobility of scientists. He believed that this structure should be supported by a competitively funded teaching and research environment within and outside government. He argued that research should be ‘applicable’, without specifying the time frame too closely and he argued most persuasively that it was self-defeating for society or government to erode the autonomy of the scientific community. Notions of ‘basic’ and ‘applied’ research and the needs of ‘stakeholders’ were cornerstones of John’s philosophy long before scientific institutions began to draw these distinctions. The Whitlam government that had commissioned the report was dismissed before it saw the light of day so political reaction was never tested. Nevertheless, many of its recommendations have been realized in principle although John’s hopes for quality control might have been disappointed.
John’s three years as foundation Director of the CSIRO Institute of Physical Sciences from 1979 were energetic and idealistic and Institute meetings were wonderful forums for interdisciplinary discussion among ten Divisional Chiefs of wide-ranging persuasion. John’s aspirations, based on scientific quality, were evident in the four major divisional reviews he conducted, although his impatience with so-called Standards measurement raised disquiet among staff who felt that he did not understand the challenge of fine measurement. His aspirations were also challenged by the ambitions of fellow Directors and some Chiefs who did not share his vision for CSIRO as set out in the Task Force Report. The ultimate factor, which he deplored, arose, however, from political pressure to use ‘external income’ as a measure of scientific achievement and for CSIRO thence to operate, as former Minister for Science Barry Jones put it, like an upbeat panel-beating shop.
John’s contributions to the Australian Academy of Science were similarly energetic. He was proud of the distinction that his Fellowship was considered by four of the six Section Committees of the Academy; those of mathematical, physical, terrestrial and biological sciences. He served on the Council from 1972 to 1978. In Council meetings John’s contributions were wide-ranging, witty and irreverent in presentation but balanced in their recommendations. He was an activist Secretary (Biological Sciences) from 1974 to 1978, and played a major role in planning the intellectual (as opposed to the ceremonial) activities of the Academy’s 25th anniversary in 1978. As part of those activities, he introduced the symposium ‘Science and the polity: Ideals, illusions and realities’ and contributed significantly to discussion of issues of scientific accountability and autonomy that featured in the Science Task Force report.
John Philip was an enthusiastic traveller, a connoisseur of architecture, a catholic reader and a published poet. He loved cooking and eating and he was a charming host and vivacious dinner guest. He played chess with a computer, watched sport on television and admired and loved his cats.
His passion for literature developed at Scotch College in what he describes as the magnificent library created when Wesley and Scotch Colleges combined during the Second World War. His favourite subject was mathematics but his most rewarding achievement, in his opinion, was his period as Assistant Editor of the Scotch Collegian in 1942. His interest in poetry and literature sustained him during his engineering course and continued throughout his life. His first poems were published in 1943 when he was 16 years old. Subsequent poems appeared regularly in Australian Poetry, in Overland, in Quadrant and in at least four Australian anthologies, the most recent of which was The New Oxford Book of Australian Verse edited by Les Murray. The poet and the scientist in John tended to be quite distinct. A note, for example, in an anthology of Australian poetry edited by Inglis Moore focused on his poetry but conceded that ‘He also contributes papers to scientific journals’. An obituary of Philip Jones (1999) observed that John was ‘a rare creature of two cultures’ but most of his scientific and artistic friends were brought together for the first time at his grave.
John’s architectural interests appeared with the design of the Philip home in Canberra, one of three complementary houses in Vasey Crescent, Campbell. Frances conceived the design and John and the architect, Sir Roy Grounds, supported her ideas. In 1998 the triptych won a coveted 25 Year Award from the Royal Australian Institute of Architects. John’s next architectural experience was CSIRO’s Pye Laboratory. This was made possible by a bequest from Fred Pye, a New South Wales grazier, and was to be John’s workplace for thirty years. In this case Frances’ vision of an airy building with offices looking down into a naturally lit native garden courtyard surrounded by glass-walled laboratories was developed by John and realised by architect Ken Woolley. Australian Department of Works bureaucrats were unwilling to accept that so elegant and functional a building might be constructed at so little cost and resisted its open design, so cleverly planned to promote interaction between the occupants. It was always considered a joy and a privilege to work there.
These architectural interests resulted in John’s lay membership of the committee that awarded the Sulman Prize for Architecture and his fascination for design much strengthened the Australian Academy of Science’s Precinct and House Committee.
Personally, John Philip was competitive and self-opinionated with rigorous standards of academic excellence that he also expected of his fellows. Early in his career he was known to be ‘difficult’ and Fred (F. W. G.) White, the then Deputy Chairman of CSIRO, delicately observed in 1957 that ‘Philip’s personality does not attract everyone’. Otto Frankel noted at the same time that ‘He seems to thrive in an environment where he is an intellectual king pin in a machine which is not quite as alive to his own way of thinking as he is himself’. John’s use of personal hyperbole in argument was understood and even appreciated by some of his colleagues but it antagonized many more and he comprehensively and frequently failed his own much-repeated aphorism, that ‘it is unforgivable to be rude by accident’.
Difficulties in the early days of computers also made him, forever, impatient of their use and he relied on an electro-mechanical Monroe calculator, called Marilyn, for much of his career. When a problem really attracted his attention, he worked at a prodigious rate, generally lying, lightly clad, on the floor. He tabulated all his results often to seven places of decimals and he drew graphs by hand. He used the computational skill of others, particularly John Knight, later in his career.
A staff survey in the mid-1980s rated John second to none as a scientist and second-last as a manager of people. He claimed that he never wanted to be loved, but the latter ranking cut deep and it was unfair. In particular, his discriminating appreciation of excellent data, and his wholehearted support for skilled experimentalists more than compensated for what he recognised as his own clumsiness. CSIRO managers did not love him much either. His election to the Royal Society in 1974, for example, was not widely publicized by the organization because, as the Chairman said in response to criticism from Bill (C. H. B.) Priestley, ‘While a Fellowship of the Royal Society is of very great significance to us, it does not mean much to the average newspaper reader and consequently is only of minor interest to the press.’
At the same time, John could be very kind, although he concealed his mothering of a collection of acquaintances. His devoted care for his aged father and his deep and abiding gratitude and love for his wife of fifty years, Frances, were also private.
Frances, their children Peregrine, Julian and Candida and others of his family and friends buried him near his parents in a small graveyard at Franklinford, near Castlemaine, Victoria, on 9 July 1999. His last poem, published in Quadrant in 1998, is inscribed on his grave:
Indigestible in life, indeed obtuse,
prone to argument and even clashes
my body please do not reduce
to an old tin of greasy ashes
But take a measure less obscene:
throw my remnants in the earth,
let worms and microbes pick me clean,
angular, indigestible, as at birth
This memoir was originally published in Historical Records of Australian Science, vol.16, no.2, 2005. It was written by David Smiles, CSIRO Land and Water, Canberra, Australia.
Numbers in brackets refer to the bibliography.
Access to CSIRO and National Library of Australia archives is gratefully acknowledged, as are observations by Drs Frank Bradley, Tom Denmead, David Elrick, Lloyd Evans, Phillip Ford and Ian White and by Dr Jim Mitchell, Co-Archivist of Scotch College, Melbourne. I include, almost verbatim, Peter Raats’ perceptions of John Philip as a mathematician. John’s wife, Frances, and his family made personal contributions.
© 2018 Australian Academy of Science