• 지능정보연구
• /
• 제18권4호
• /
• pp.117-127
• /
• 2012

최근 온라인 소셜 네트워크 서비스(SNS)의 사용자가 크게 늘어나고 있으며 다양한 분야에서 SNS의 사용자 관계 구조 및 메시지를 분석하기 위한 연구를 진행하고 있다. 그러나 대부분의 소셜 네트워크 분석 방법들은 노드 사이의 최단 거리를 기초로 하고 있으므로 계산 시간이 오래 걸린다. 이는 점차 대형화 되어가는 SNS의 데이터를 여러 분야에서 활용하는데 걸림돌이 되고 있다. 이에 따라 본 논문에서는 SNS의 사용자 그래프에서 사용자간 최단거리를 빠르게 찾기 위한 휴리스틱 기반의 최단 경로 탐색 방법을 제안한다. 제안하는 방법은 1) 트리로 표현된 소셜 네트워크에서 시작 노드와 목표 노드를 설정한다. 그리고 2) 만약 목표 노드가 경사 트리의 단말에 있다면 경사 트리가 시작하는 노드를 임시 골 노드로 설정한다. 마지막으로 3) 연결의 차수를 평가값으로 하는 휴리스틱 기반 최단거리 탐색을 수행한다. 이렇게 최단거리를 탐색한 후 매개 중심성 분석(Betweenness Centrality) 및 근접 중심성(Closeness Centrality)를 계산한다. 제안하는 방법을 사용하면 소셜 네트워크 분석에서 가장 많은 시간이 필요한 최단거리 탐색을 빠르게 수행할 수 있으므로 소셜 네트워크 분석의 효율성을 기대할 수 있다. 본 논문에서 제안하는 방법을 검증하기 위하여 약 16만 명으로 구성된 SNS에서의 실제 데이터를 이용하여 매개 중심성 분석과 근접 중심성 분석을 수행하였다. 실험 결과, 제안하는 방법은 전통적 방식에 비하여 매개 중심성, 근접 중심성의 계산 시간이 각각 6.8배, 1.8배 더 빠른 결과를 보였다. 본 논문에서 제안한 방법은 소셜 네트워크 분석의 시간을 향상시켜 여러 분야에서 사회 현상 및 동향을 분석하는데 유용하게 활용될 수 있다.

• 자연과학논문집
• /
• 제1권
• /
• pp.61-113
• /
• 1987

The increasing emphasis placed on the production of fine turf for lawns, golf courses, parks, and other recreational sites has led to many unsolved problems as to how such turf could be best established and mainteined. For this purpose, a series of experiments were conducted under con ditions of pot and field. The results obtained were as follows EXPERIMENT I. The effect of nitrogen fertilizer and clipping interval on Zoysia japonica. 1. Increasing the rate of nitrogen and frequent clipping increased tiller number of Zoysis japonica and the maximum number of tillers were obtained from 700 kg N application and freqnent clippings (10 days interval ) in October. Treatment of 350kg N with 10 days clipping interval increased tillers much more than those of 700 kgN with 20 and 30 days clipping intervals. 2. The average number of green leaves occurred during the growth period maximized by applying 700 kg N and clipping 10 days interval. 3. Increasing tiller numbers significantly decreased tops DM weight per tiller by clipping plants at interval of 10 and 20 days, irrespective of nitrogen applied, and with nil N, at the interval of 30 days. By applying 700 kg N, however, top DM weight per tiller increased as the number of tillers increased consistently. 4. The highest top DM weight was achieved from late August to early September by applying 350 and 700kgN. 5. During the growth period, differences in unders ( stolon + root ) DM weight occurred bynitrogen application were found between nil N and two applied nitrogen levels, whereas, at the same level of nitrogen applied, the increase in stolon DM weight enhanced by lengthening the clipping interval to 30 days. 6. Nitrogen efficiency to green leaves, stolon nodes and DM weight of root with high nitrogen was achieved as clipping interval was shortened. 7. By increasing fertilizer nitrogen rate applied, N content n the leaves and stems of Zoysiajaponica was increased. On the other hand, N content in root and stolon had little effect onfertilizer nitrogen, resulting in the lowest content among plant fractions. The largest content of N was recorded in leaves. Lengthening the clipping interval from 10 or 20 to 30 days tends to decrease the N content in the leaves and stems, whereas this trend did not appeared in stolon androot. 8. A positive correlations between N and K contents in tops and stolon were established andthus K content increased as N content in tops and stolon increased. Meanwhile, P content was not affected by N and clipping treatments. 9. Total soluble carbohydrate content in Zoysia japonica was largest in stolon and stem, and was reduced by increasing fertilizer nitrogen rate. Reduction in total soluble carbohydrate due to increased nitrogen rate was severer in the stolons and stems than in the leaves. 10. Increasing the rate of nitrogen applied increased the number of small and large vascular bundles in leaf blade, but shortened distance among the large vascular bundles. Shortening the clipping interval resulted in increase of the number of large vascular bundles but decrease ofdistance between large vascular bundles.EXPERIMENT II. Growth response of Zoysia japonica imposed by different plant densities. 1. Tiller numbers per unit area increased as plant density heightened. Differences in num ber between densities at higher densities than 120 D were of no significance. 2. Tiller numbers per clone attained by 110 days after transplanting were 126 at 40D,77 at 80D, 67 at 120D, 54 at 160D, and 41 at 200D. A decreasing trend of tiller numbers per clone with increasing density was noticable from 100 days after transplanting onwards. 3. During the growth period, the greatest number of green leaves per unit area were attainedin 90days after transplanting at 160D and 200D, and 100 days after transplanting at 40D, 80Dand 120D. Thus the period to reach the maximum green leaf number with the high plantdensity was likely to be earlier that with the low plant density. 4. Stolon growth up to 80 days after transplaning was relatively slow, but from 80 daysonwards, the growth quickened to range from 1.9 m/clone at 40D to 0.6m/clone at 200Din 200 days after transplanting, these followed by the stolon node produced. 5. Plant density did not affect stolon weight/clone and root weight/clone until 80 daysafter transplanting. 6. DM weight of root was heavier in the early period of growth than that of stolon, butthis trend was reversed in the late period of growth : DM weight of stolon was much higherthan that of root.EXPERIMENT Ill. Vegetative growth of Zoysia japonica and Zoysia matrella as affected by nitrogen and clipping height. 1. When no nitrogen was applied to Zoysia japonica, leaf blade which appeared during theAugust-early September period remained green for a perid of about 10 weeks and even leavesemerged in rate September lived for 42 days. However, leaf longevity did not exceed 8 weeks asnitrogen was applied. In contrast the leaf longevity of Zoysia matrella which emerged during the mid August-earlySeptember period was 11 weeks and, under the nitrogen applied, 9 weeks, indicating that thelife-spen of individual leaf of Zoysia matrella may be longer than that of Zoysia japorica. Clipping height had no effect on the leaf longevity in both grasses. 2. During the July-August period, tiller number, green leaf number and DM weightof Zoysia japonica were increased significantly with fertilizer nitrogen, but were not with twolevel of clipping height. This trend was reversed after late September ; no effect of nitrogen wasappeared. Instead, lax clipping increased tiller number, green leaf number and DM weight. Greenleaves stimulated by lax clipping resulted in the occurrance of more dead leaves in late October. 3. Among the stolons outgrown until early September, the primary stolon was not influencedby nitrogen and clipping treatments to produce only 2-3 stolons. However, 1st branch stoIon asaffected by nitrogen increased significantly, so most of stolons which occurred consisted of 1st branch stolons. 4. Until early September, stolon length obtained at nil nitrogen level was chiefly caused bythe primary stolons. By applying nitrogen, the primary stolons of Zoysia japonica waslonger than 1st branch stolons when severe clipping was involved and in turn, shorter than 1stbranch stolons when lax clipping was concerned. In Zoysia matrella, 1st branch stolons were muchlonger than the primary stolon when turf was clipped severely but in conditions of lax clippingthere was little difference in length between primary and 1st branch stolons. 5. Stolon nodes of both Zoysia japonica and Z. matrella were positively influenced by nit rogen, but no particular increase by imposing clipping height treatment was marked in Zoysiamatrella. Although the stolon of Zoysia japonica grew until late October, the growthstimulated by nitrogen was not so remarkable as to exceed that by nil N.

• 마케팅과학연구
• /
• 제19권4호
• /
• pp.32-43
• /
• 2009

信任在心理学, 经济学, 社会学中已被广泛研究, 其重要性不仅在市场营销中被强调, 在一般商业原则中也被强调. 供应商和买家之间的关系与过去不同, 过去的关系需要相当大的私人网络优势, 并可能涉及不道德的商业行为. 而在以工业营销成功的为核心的二十一世纪激烈的全球竞争中, 供应商和买家之间的关系是伙伴关系. 在相互合作的高级别信任的基础上, 通过交换的关系, 这会给买家和供应商带来长期的利益, 竞争力增强和交易成本的降低以及其他福利. 尽管现有的研究有信任的重要性, 但是在购买与供应关系中却忽视了信任的作用, 也没有系统地分析信任对关系的影响. 因此, 深入研究, 确定买家和商业服务供应商之间信任和关系绩效之间的联系是绝对需要的. 本研究中的商业服务, 包括那些支持制造业, 正作为下一代经济增长的引擎而吸引着人们的注意. 韩国政府已选择其作为制造业发展的战略领域. 由于商业服务开放市场的需求日趋激烈, 商业服务业的竞争力应该比以往得到更多的提倡. 本研究的目的是探索相互信任对买家和供应商之间的关系绩效的影响. 具体来说, 本研究在商业服务交易中提出了一个关于信任-关系绩效的理论模型, 并实证检验根据模型而提出的假设. 这项研究表明, 研究结果有战略意义. 本研究通过多种方法收集经验数据. 这些方法包括通过电话, 邮件和面试. 作为样本的公司是在韩国供应和购买商业服务的以知识为本的公司. 本研究收集的是二进的基础数据. 每个样本公司对包括购买公司及其相应的供应公司. 并跟踪调查每个公司对的相互信任. 本研究为商业服务的买卖双方提出了信任-关系绩效的模型. 该模型由信任和它的前因和后果. 买家的信任分为对供应公司的信任和对销售人员的信任. 根据Doney 和Cannon (1997)的研究我们在个人水平和组织水平上观察信任. 通常情况下, 买方是信任的受体, 但这项研究我们建议以供应商为观察受体. 因此, 它独特的关注了双边角度的知觉风险. 换言之, 供应商和买家一样, 是信任的主体, 因为交易通常是双边的. 从这个角度来看, 供应商对买家信任和买方对供货商的信赖一样重要. 供应商的信任从某种程度上受它信任的买方公司和买家的影响. 这种使用个人水平和组织水平的信任分类是根据Doney 和Cannon (1997)的研究. 信任影响供应商的选择, 这是一项双向放的工作. 供应商们积极参与供应商选择过程中, 和买家密切的一起工作. 此外, 该过程从某种程度上受每一方信任的合作伙伴的影响. 挑选过程包括一些步骤: 识别, 信息检索, 供应商选择和绩效评价. 作为这一进程的结果, 买家和供应商都进行绩效评估, 并就这些结果为基础, 采取有形或无形的纠正行动. 本研究中使用的关于信任的测量问项是根据Mayer, Davis 和 Schoorman (1995) 以及Mayer和Davis (1999)的研究发展起来的. 根据他们的建议, 有关信任的三个方面的研究包括有能力, 善和完整. 根据商业服务这个背景我们调整了原来的问题. 例如, 如 "他/她的专业能力" 已被改为 "当我们讨论我们的产品时销售人员表现出专业能力. "这项研究使用的测量问项不同于在以往的研究中使用的问项（Rotter 1967; Sullivan和Peterson 1982; Dwyer和Oh 1987. 本研究中有关信任的前因后果的测量问项是根据Doney和Cannon (1997)的研究为基础制定的. 根据商业服务这个背景我们调整了原来的问题. 特别是, 问题被设计为对买家和供应商以解决下列因素: 信誉 (诚信, 客户服务, 良好意愿), 市场地位 (公司规模, 市场份额, 在行业中的地位), 愿意定制（产品, 过程, 交付), 信息共享(专有信息, 个人信息), 愿意保持良好关系, 认为专业, 权威授权, 买方与卖方的相似性, 以及接触频率. 作为信任相应的变量, 我们对关系绩效进行了测试. 关系绩效分为有形的影响, 无形影响, 和副作用. 有形的影响包括财务业绩;无形的影响, 包括关系的改善, 网络开发, 以及内部员工的满意度;副作用包括既不是有形影响也不是无形影响的影响. 我们联系了350对公司, 105对公司答复了我们. 由于不完整我们删除了5对公司, 105对公司被用于数据分析. 用于数据分析的回应率为30%(三百五十零分之一百零五), 高于工业营销的平均回复比率. 至于回复的公司的特点, 大多数的公司运作的商业服务既为买方(85.4%)也为供应商(81.8%). 大部分买家是做消费品贸易(76%), 而供应商的大部分(70%)是做工业品贸易. 这可能意味着买家的过程是购入材料, 部件和组件从而生产消费品成品. 正如他们对他们与合作伙伴关系的长度的报告表示, 供应商比买家有更长的商业关系. 假设1测试买方-供应方特点对信任的影响. 销售人员的专业度(t=2.070, p<0.05)和权威授权(t=2.328, p<0.05)积极影响买方对供应方的信任. 另一方面, 权威授权(t=2.192, p<0.05)积极影响供应方对买方的信任. 对买方和供应方来说, 权威授权的程度对保持对彼此的信任有关键作用. 假设2测试买卖双方关系特点对信任的影响. 买家倾向于信任供应方, 因为供应方总是尽全力联系买方(t=2.212, p<0.05)这种倾向性在供应方方面也表现得很强(t=2.591, p<0.01). 另一方面, 供应商对买方的信任是由于供应商感知买家与自己的相似性(t=2.702, p<0.01). 这一发现证实了Crosby, Evans, 和Cowles(1990)的研究结果. 他们的结果表明供应方和买方通过商务或私务的定期会议来建立彼此的联系. 假设3测试信任对感知风险的影响. 结果表明无论对买方还是供应方, 信任越低, 感知风险就越大(买方: t =-6.621, p<0.01; 供应方: t=-2.437, p<0.05). 有趣的是, 这一趋势已被证明对买方更强. 这种较高水平的感知风险的一个可能的解释是在商业服务交易中买方通常比供应方感知到更大的风险. 为此, 有必要对供应商对买方实施减少风险的战略. 假设4测试信任对信息搜集. 根据结果, 对供应方和买方, 与预期相反, 信任取决于他们合作伙伴的名誉(买方t=2.929, p<0.01; 供应方t=2.711, p<0.05). 这一发现表明, 具有良好信誉的供应商往往是可信的. 以往的经验并没有显示出任何与买家或供应商信任的重要关系. 假设5测试信任对供应方/买方选择的影响. 与买方不同, 当供应方认为以往与买方的交易重要时, 供应方倾向信任买方(t=2.913 p<0.01). 但是, 本研究并没有现实资源忠诚和买方对供应方的信任之间有显著关系. 假设6测试的是信任对关系绩效的影响. 对买方和供应方, 当财务表现被报告提高时, 他们比较信任他们的合作伙伴(买方: t=2.301, p<0.05;供应方: t=3.692, p<0.01). 有趣的是, 这种趋势在供应方比较明显. 类似的, 当竞争力被报告提高时, 买卖双方比较信任他们的合作伙伴(买方t=3.563, p<0.01 ; 供应方t=3.042, p<0.01). 对供应方来说, 当对买方信任时效率和生产力会提高(t=2.673, p<0.01). 其他绩效指标与信任没有显著关系. 这项研究结果有一定的战略意义. 首先和最重要的是, 以信任为基础的交易对供应商和买家而言都是有益的. 根据研究证实, 通过努力建立和保持相互信任可以使财务表现提高. 同样, 可以通过同样的努力提高竞争力. 第二, 以信任为基础的交易能够减少购买情况中的感知风险. 这对供应商和买家都有启示. 人们普遍认为, 在一个高度参与的采购情况中买家感知到更高的风险. 为了减少风险, 以往的研究已建议供应商制定降低风险的策略. 而本研究的特点是从双边角度关注知觉风险. 换言之, 供应商也容易存在风险, 特别是当他们提供的服务, 需要非常先进的技术, 操作和维护. 因此, 购买者和供应商必须一起密切合作解决问题. 因此, 相互信任在问题解决过程中起着关键作用. 第三, 在这项研究中发现, 销售人员有更多的授权, 他或她越被信任. 这一发现从战术角度看是非常重要的. 建立信任是一个长期的任务, 然而, 当互信尚未开发, 供应商能够通过授权销售人员做出某些决定来克服遇到的问题, 这一结论也适用于供应商.

• 생물환경조절학회지
• /
• 제5권2호
• /
• pp.215-235
• /
• 1996

본 연구에서 수행한 Model 시뮬레이션에 의한 열환경 분석 기법은 지역별로 다양한 기상여건 하에서 대상온실의 난방 및 냉방부하를 보다 합리적으로 예측할 수 있을 뿐만 아니라 냉방이나 난방용 시스템의 결정을 비롯한 난방대책을 수립하고, 에너지 이용 전략의 수립이나 계절적인 작부계획 수립, 온실산업용 적지선정 등에 유익하게 활용될 수 있을 것이라 판단된다. 본 연구에서는 온실의 적극적인 환경조절 유형을 난방과 냉방의 두 가지로 대별하고, 난방 소요열량 산정을 비롯하여 야간의 보온 커튼효과, Heating Degree-Hour 산정 등 난방과 관련된 시뮬레이션은 동적 모형을 이용하여 시간별, 일별 및 월별로 검토하였으며, 환기를 비롯한 차광, 증발냉각시스템의 효과 분석은 정적모형을 이용하여 검토하였다. 특히 하절기 지하수와 같은 저온수를 직접 이용하거나 Heat Pump를 통하여 확보될 수 있는 저온수를 이용하여 온실의 피복면에 살수함으로서 확보할 수 있는 온실냉방효과를 검토하는 데는 1.2m$\times$2.4m 크기의 모형온실을 제작하여 기초실험을 수행함으로서 동절기의 수막시스템의 보온효과와 마찬가지로 하절기 냉방 효과를 거둘 수 있다는 가능성을 확인하였다. 본 연구에 활용된 온실의 수치 환경모형 중 난방관련 시뮬레이션용 동적 수치모형은 소기의 목적을 달성하는데 충분히 응용될 수 있는 이론모형이다. 이 이론모형이 범용성이 높은 것은 온실 내ㆍ외의 미기상 변화, 특히 난방이나 냉방이 본격적으로 요구되는 기간동안에 온도, 습도, 일사, 풍속 등의 미기상 인자들을 면밀하게 관찰하여 실측된 자료를 바탕으로 개발되었고, 다양한 자료에 의해 충분히 검정되었기 때문이다. 본 연구에서는 경남 진주지역의 어느 특정 기간(1987년)의 시간별 기상자료를 중심으로 온실의 열적 환경변화에 대한 수치모형 시뮬레이션을 실시하였으며, 아직 수치모형에 의한 시뮬레이션이 불가능한 일부 냉방효과를 검토하는 데는 모형 실험을 실시하였으며, 그 결과를 요약하면 다음과 같다. 1. 주간과 야간의 설정온도를 달리하고 다단계 변온조절방식으로 시뮬레이션을 행한 결과 난방 소요열량은 난방 설정온도에 따라 현저한 차이를 보였다. 특히 주간 설정온도에 비하여 야간 설정온도가 난방 소요열량에 예민하게 영향을 미치므로 야간의 설정온도 결정에 신중을 기해야 할 것으로 판단된다. 2. 기존의 Heating Degree-Hour 자료는 평균 외기온을 중심으로 임의의 설정온도에 대하여 산정된 값이므로 난방 소요열량에 대한 상대적인 비교수단은 되나 고려되는 기상인자의 제한과 설정온도의 임의성 때문에 실용성이 부족하다. 따라서 본 연구에서 제시된 것처럼 온실 주변의 제반 미기상 인자나 경계조건이 반영됨은 물론 작물의 생육상태 및 구체적인 설정온도까지도 고려하는 동적 수치모형으로 시시각각으로 예측된 실내기온을 중심으로 재배기간 동안의 난방열량을 적산함이 합리적이라 판단된다. 기존의 MDH 자료로 난방 설계를 할 경우에는 지나치게 과잉설계 될 가능성이 있다. 3. 산정된 난방 소요열량은 물론 커튼의 보온성능도 월별 기상여건에 따라 현저한 차이를 보이며, 시뮬레이션에 이용된 커튼의 경우 높은 보온효과를 보임으로서 년 평균 50% 이상의 난방 에너지를 절감할 수 있으며, 동절기 3-4개월의 집중 난방기에 에너지가 크게 절감됨을 발견할 수 있다. 4. 고온기 환기성능은 온실의 구조, 기상조건, 작물의 생육상태 등에 따라 다소의 차이가 있으나 환기율에 의해 크게 좌우되며, 시뮬레이션에 이용된 두 가지 농가보급형 온실 모두 환기율의 증가에 따른 실내기온의 강하 효과가 환기율이 1회/min 정도를 넘어서면서 급격히 둔화되는 현상을 보인다. 이는 기존에 권장되고 있는 적정 환기율인 1회/min 전후의 환기 시스템을 갖추는 것이 합리적임을 확인해 준다. 5. 작물이 성숙된 유리온실에서 외기의 상대습도가 50%인 쾌청한 주간동안 연속적으로 1회/min로 환기를 시킬 경우 실내기온 36.5$^{\circ}C$의 대조구에 비한 온도강하는 50% 차광만 했을 시 2.6$^{\circ}C$이고 효율 80%의 Pad ＆ Fan 시스템만 작동시 6.1$^{\circ}C$ 정도이며, 차광과 냉각시스템을 동시에 작동시는 약 8.6$^{\circ}C$로서 외기온보다 3.3$^{\circ}C$가 낮은 28$^{\circ}C$까지 실내온도를 낮출 수 있으나, 동일 조건하에서 외기의 상대습도가 80%로 높은 경우에는 Pad ＆ Fan시스템에 의한 온도강하가 2.4$^{\circ}C$에 불과하여 50% 차광하에서도 외기온 이하로 실내온도를 낮출 수 없음을 알 수 있다. 6. 하절기 3개월(6/1-8/31)동안 Pad ＆ Fan 시스템의 냉방효과($\Delta$T)는 설정된 작동 온도에 따라 다소 차이를 보일 것으로 예상되나 본 시뮬레이션에서 설정한 시스템의 작동 온도 27$^{\circ}C$에서 상대습도와의 상관관계는 대략 다음과 같았다: $\Delta$T= -0.077RH＋7.7 7. 전형적인 하절기 주간기상 하에서 경시적 냉방효과를 분석한 결과 환기만으로는 실내기온을 외기온 보다 5$^{\circ}C$ 높게 유지하는 정도가 고작이고, 차광이나 증발식 냉방시스템 만으로는 작물이 성숙한 단계에서조차도 외기온 이하로 떨어뜨리기가 어려우나 차광과 아울러 증발식 냉방을 병행할 경우에는 작물상태에 따라 다소 차이는 있지만 실내기온을 외기온보다 2.0-2.3$^{\circ}C$ 낮게 유지할 수 있음을 발견할 수 있다. 8. 일사가 차단된 27.5-28.5$^{\circ}C$의 외기온하에서 6.5-8.5$^{\circ}C$의 냉수를 온실 바닥면적 1$m^2$당 1.3 liter/min의 유량으로 온실표면에 살수했을 때 실내기온을 외기온보다 1$0^{\circ}C$ 낮은 16.5-18.$0^{\circ}C$ 정도로 낮출 수 있었다. 앞으로 살수 수온(T$_{w}$ )이나 외기온(T$_{o}$ ) 뿐만아니라 살수율(Q)에 따라 온실기온 (T$_{g}$ )에 미치는 상관 관계 T$_{g}$ = f(T$_{w}$ , Q, T$_{o}$ )를 구명하여 지하수 자체 또는 Heat Pump를 이용한 지하수온 이하의 냉수로 온실냉방의 가능성을 구명하는 것이 앞으로의 과제이다.

• 한국농공학회지
• /
• 제16권1호
• /
• pp.3225-3262
• /
• 1974

The purposes of this thesis are to clarify experimentally the variation of ground water temperature in tube wells during the irrigation period of paddy rice, and the effect of ground water irrigation on the growth, grain yield and yield components of the rice plant, and, furthermore, when and why the plant is most liable to be damaged by ground water, and also to find out the effective ground water irrigation methods. The results obtained in this experiment are as follows; 1. The temperature of ground water in tube wells varies according to the location, year, and the depth of the well. The average temperatures of ground water in a tubewells, 6.3m, 8.0m deep are $14.5^{\circ}C$ and $13.1^{\circ}C$, respercively, during the irrigation period of paddy rice (From the middle of June to the end of September). In the former the temperature rises continuously from $12.3^{\circ}C$ to 16.4$^{\circ}C$ and in the latter from $12.4^{\circ}C$ to $13.8^{\circ}C$ during the same period. These temperatures are approximately the same value as the estimated temperatures. The temperature difference between the ground water and the surface water is approximately $11^{\circ}C$. 2. The results obtained from the analysis of the water quality of the "Seoho" reservoir and that of water from the tube well show that the pH values of the ground water and the surface water are 6.35 and 6.00, respectively, and inorganic components such as N, PO4, Na, Cl, SiO2 and Ca are contained more in the ground water than in the surface water while K, SO4, Fe and Mg are contained less in the ground water. 3. The response of growth, yield and yield components of paddy rice to ground water irrigation are as follows; (l) Using ground water irrigation during the watered rice nursery period(seeding date: 30 April, 1970), the chracteristics of a young rice plant, such as plant height, number of leaves, and number of tillers are inferior to those of young rice plants irrigated with surface water during the same period. (2) In cases where ground water and surface water are supplied separately by the gravity flow method, it is found that ground water irrigation to the rice plant delays the stage at which there is a maximum increase in the number of tillers by 6 days. (3) At the tillering stage of rice plant just after transplanting, the effect of ground water irrigation on the increase in the number of tillers is better, compared with the method of supplying surface water throughout the whole irrigation period. Conversely, the number of tillers is decreased by ground water irrigation at the reproductive stage. Plant height is extremely restrained by ground water irrigation. (4) Heading date is clearly delayed by the ground water irrigation when it is practised during the growth stages or at the reproductive stage only. (5) The heading date of rice plants is slightly delayed by irrigation with the gravity flow method as compared with the standing water method. (6) The response of yield and of yield components of rice to ground water irrigation are as follows: \circled1 When ground water irrigation is practised during the growth stages and the reproductive stage, the culm length of the rice plant is reduced by 11 percent and 8 percent, respectively, when compared with the surface water irrigation used throughout all the growth stages. \circled2 Panicle length is found to be the longest on the test plot in which ground water irrigation is practised at the tillering stage. A similar tendency as that seen in the culm length is observed on other test plots. \circled3 The number of panicles is found to be the least on the plot in which ground water irrigation is practised by the gravity flow method throughout all the growth stages of the rice plant. No significant difference is found between the other plots. \circled4 The number of spikelets per panicle at the various stages of rice growth at which_ surface or ground water is supplied by gravity flow method are as follows; surface water at all growth stages‥‥‥‥‥ 98.5. Ground water at all growth stages‥‥‥‥‥‥62.2 Ground water at the tillering stage‥‥‥‥‥ 82.6. Ground water at the reproductive stage ‥‥‥‥‥ 74.1. \circled5 Ripening percentage is about 70 percent on the test plot in which ground water irrigation is practised during all the growth stages and at the tillering stage only. However, when ground water irrigation is practised, at the reproductive stage, the ripening percentage is reduced to 50 percent. This means that 20 percent reduction in the ripening percentage by using ground water irrigation at the reproductive stage. \circled6 The weight of 1,000 kernels is found to show a similar tendency as in the case of ripening percentage i. e. the ground water irrigation during all the growth stages and at the reproductive stage results in a decreased weight of the 1,000 kernels. \circled7 The yield of brown rice from the various treatments are as follows; Gravity flow; Surface water at all growth stages‥‥‥‥‥‥514kg/10a. Ground water at all growth stages‥‥‥‥‥‥428kg/10a. Ground water at the reproductive stage‥‥‥‥‥‥430kg/10a. Standing water; Surface water at all growh stages‥‥‥‥‥‥556kg/10a. Ground water at all growth stages‥‥‥‥‥‥441kg/10a. Ground water at the reproductive stage‥‥‥‥‥‥450kg/10a. The above figures show that ground water irrigation by the gravity flow and by the standing water method during all the growth stages resulted in an 18 percent and a 21 percent decrease in the yield of brown rice, respectively, when compared with surface water irrigation. Also ground water irrigation by gravity flow and by standing water resulted in respective decreases in yield of 16 percent and 19 percent, compared with the surface irrigation method. 4. Results obtained from the experiments on the improvement of ground water irrigation efficiency to paddy rice are as follows; (1) When the standing water irrigation with surface water is practised, the daily average water temperature in a paddy field is 25.2$^{\circ}C$, but, when the gravity flow method is practised with the same irrigation water, the daily average water temperature is 24.5$^{\circ}C$. This means that the former is 0.7$^{\circ}C$ higher than the latter. On the other hand, when ground water is used, the daily water temperatures in a paddy field are respectively 21.$0^{\circ}C$ and 19.3$^{\circ}C$ by practising standing water and the gravity flow method. It can be seen that the former is approximately 1.$0^{\circ}C$ higher than the latter. (2) When the non-water-logged cultivation is practised, the yield of brown rice is 516.3kg/10a, while the yield of brown rice from ground water irrigation plot throughout the whole irrigation period and surface water irrigation plot are 446.3kg/10a and 556.4kg/10a, respectivelely. This means that there is no significant difference in yields between surface water irrigation practice and non-water-logged cultivation, and also means that non-water-logged cultivation results in a 12.6 percent increase in yield compared with the yield from the ground water irrigation plot. (3) The black and white coloring on the inside surface of the water warming ponds has no substantial effect on the temperature of the water. The average daily water temperatures of the various water warming ponds, having different depths, are expressed as Y=aX+b, while the daily average water temperatures at various depths in a water warming pond are expressed as Y=a(b)x (where Y: the daily average water temperature, a,b: constants depending on the type of water warming pond, X; water depth). As the depth of water warning pond is increased, the diurnal difference of the highest and the lowest water temperature is decreased, and also, the time at which the highest water temperature occurs, is delayed. (4) The degree of warming by using a polyethylene tube, 100m in length and 10cm in diameter, is 4~9$^{\circ}C$. Heat exchange rate of a polyethylene tube is 1.5 times higher than that or a water warming channel. The following equation expresses the water warming mechanism of a polyethylene tube where distance from the tube inlet, time in day and several climatic factors are given: {{{{ theta omega (dwt)= { a}_{0 } (1-e- { x} over { PHI v })+ { 2} atop { SUM from { { n}=1} { { a}_{n } } over { SQRT { 1+ {( n omega PHI) }^{2 } } } } LEFT { sin(n omega t+ { b}_{n }+ { tan}^{-1 }n omega PHI )-e- { x} over { PHI v }sin(n omega LEFT ( t- { x} over {v } RIGHT ) + { b}_{n }+ { tan}^{-1 }n omega PHI ) RIGHT } +e- { x} over { PHI v } theta i}}}}{{{{ { theta }_{$\infty$ }(t)= { { alpha theta }_{a }+ { theta }_{ w'} +(S- { B}_{s } ) { U}_{w } } over { beta } , PHI = { { cpDU}_{ omega } } over {4 beta } }}}} where $\theta$$\omega$; discharged water temperature($^{\circ}C$) $\theta$a; air temperature ($^{\circ}C$) $\theta$$\omega$';ponded water temperature($^{\circ}C$) s ; net solar radiation(ly/min) t ; time(tadian) x; tube length(cm) D; diameter(cm) ao,an,bn;constants determined from $\theta$$\omega$(t) varitation. cp; heat capacity of water(cal/$^{\circ}C$ ㎥) U,Ua; overall heat transfer coefficient(cal/$^{\circ}C$ $\textrm{cm}^2$ min-1) $\omega$;1 velocity of water in a polyethylene tube(cm/min) Bs ; heat exchange rate between water and soil(ly/min)